EPA-650/2-74-047
June 1974
Environmental Protection Technology Series
I
1
SSE2
\
UJ
O
-------
DESIGN OF AN OPTIMUM DISTILLATE
OIL BURNER FOR CONTROL
OF POLLUTANT EMISSIONS
by
Ft. A. Dickerson and A. S. Okuda
Rocketdyne Division
Rockwell International
6633 Canoga Avenue
Canoga Park, California 9130h
Contract No. 68-02-0017
ROAP No. 21ADG-UU
Program Element No. lABOlU
EPA Project Officer: G. B. Martin
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 201*60
June 1971*
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABSTRACT
This report describes results from a research study of the pollution char-
acteristics of high-pressure atomizing, No. 2 distillate fuel oil burners.
The main emphasis was on optimizing burner design to minimize pollutant
emissions when these burners are fired into refractory-lined combustion
chambers. The atomization characteristics, flow profiles, and composition
profiles in the combustion zones of several commercial burners were deter-
mined experimentally. Mass median droplet diameters ranged from 60 to 90
microns for 0.50- to 1.50-gph oil spray nozzles. Nitric oxide formation
was most prevalent in the near-stoichiometric combustion zones where local
flow conditions led to vigorous mixing of gases. These data were used to
guide the design of variable geometry burners that were used to optimize
the burner geometry for minimization of pollutant emissions. The optimum
geometry burners were fabricated in fixed geometry versions and tested
extensively to verify their low air pollutant emissions. Substantial re-
ductions in NO (~50 percent) emissions were achieved by optimizing conven-
tional designs, with negligible emissions of other pollutants. Addition-
ally, several nonconventional burner designs were built and tested, two
of which led to very low nitric oxide emissions. Results from the pro-
gram have been used to develop recommendations for burner design to mini-
mize pollutant emissions.
This report was submitted in fulfillment of Contract No. 68-02-0017 by
Rocketdyne Division, Rockwell International, Canoga Park, California,
under the sponsorship of the Environmental Protection Agency. Work was
completed as of 1 February 1974.
iii
-------
CONTENTS
Page
Abstract iii
List of Figures vi
List of Tables xvi
Summary 1
Atomization Measurements 2
Tests of Commercial Burners 2
Versatile Burner Experiments 4
Optimized Fixed Geometry Burners 5
Nonconventional Burner Concepts 7
Conclusions 9
Summary of Recommended Design Practices 10
Introduction 15
Apparatus and Data Reduction 23
Atomization Test Facility 23
Air Flow Patterns 26
Oil Spray Patterns 34
Combustion Gas Flow Patterns 39
Combustion Chambers 43
Furnaces 49
Combustion Gas Composition Patterns 51
Exhaust Gas Analyses 63
Thermal Radiation 64
Nonconventional Burners 66
Oil Nozzles 75
Experimental Results 79
Atomization Measurements 80
Commercial Burner Studies 82
Interpretation of Commercial Burner Results 173
Burner Geometry Optimization Studies 188
Simulated Field Testing of Optimum Geometry Burners 207
Nonconventional Burners 218
IV
-------
CONTENTS (Continued)
Page
Conclusions 238
Optimum Burner Geometry 238
Burner Design Philosophy 238
Combustion Chamber Wall Temperature 240
Combustion Chamber Geometry 240
Appendix A
Commercial Design Practices 241
Appendix B
Factors for the Conversion of Units to the Metric System 251
-------
FIGURES
No. Page
1 Photographs of a 1-gph, Low Pollutant Emission 12
Optimized Oil Burner Head
2 Choke Diameter Versus Oil Flowrate for Minimum Nitric
Oxide Emissions, Low Smoke and Carbon Monoxide Emissions 13
3 Typical High-Pressure, Oil-Fired Gun Burner 17
4 Simplex Oil Spray Nozzle 19
5 Three Typical Air Diffuser Designs Used in the Ends
of Oil Burner Blast Tubes 19
6 Wax Flow Facility 24
7 Velocity Direction Indicator Probes 28
8 Schematic of Directional Probe Girabaling Apparatus Used
to Alter Probe Angle While Keeping the Probe Tip Stationary 29
9 Probe Gimbaling Apparatus 31
10 Interpretation of 3 Angle Curves 33
11 Location of Horizontal and Vertical Planes for
Determination of Gas Velocity Vectors 35
12 Liquid Spray Mass Flux Sampling Apparatus 37
13 Effects of Sampling Velocity on Oil Collection Rate 38
14 Oil Burner Cold-Flow Facility Schematic 40
15 Water-Cooled Probes 41
16 Schematic of Oil Burner Flame Gas Velocity Probe System 42
17 Schematic of the 8.0-Inch-Diameter Insulated Cylindrical
Combustor With Semicircular Mixing Baffles 44
18 Schematic of 30-Inch-Diameter, Hot/Cold Wall, Coaxial
Cylindrical Research Combustor for 6- to 12-gph Oil Burners 47
19 Schematic of Oil Burner Combustion Gas Sampling System 50
20 United Sensor Model GB 24-125 Gas Sampling Probe 53
21 Analytical System for Fuel Oil Burner Emission Analysis 54
22 Land 2ir Radiometer Used for Radiation Measurements 65
23 Schematic of Modified Versatile Burner With Variable Rate,
Mechanically Driven, Multiple Vane Swirler 67
VI
-------
FIGURES (Continued)
No.
24 Cross-Section View of the Fixed Geometry, 1.65-Inch-
Diameter, 30-Degree Convergent Choke 69
25 Photographs of the Mechanically Rotated, Six-Vane Intense
Swirl Burner in Various Stages of Assembly 70
26 Schematic of Displaced Oil Injection Burner Prototypes
Constructed and Hot Fired 71
27 Schematic of the Forced Recirculated Combustion Gas
Experimental Burner 73
28 Photograph of the Forced Recirculating Combustion Gas
Burner (fully instrumented) With an 8-Vane, 60-Degree
Swirler Ring, and 4 of 12 Combustion Gas Inlet Ports Plugged 74
29 Schematic of the Preheated Air Burner Test Apparatus
Using a 2-Inch-Diameter Blast Tube with an-8-Vane,
75-Degree Swirl Ring, and a 0.75-80°-C Oil Nozzle 76
30 Cross-Sectional Schematic Showing the Internal
Construction of a Typical High-Pressure Oil Atomizing1
Nozzle and the Various Spray Cone Patterns 77
31 Spray Dropsize Data Obtained Using the Method of Frozen
Wax on Various Oil Burner Spray Nozzles 81
32 Extrapolation of Delavan Data to the Mass Median Particle
Diameter Obtained for the Delavan 0.75-gph/80°-C Nozzle 83
33 Spray Dropsize Distribution for a Delavan 0.75-80°-B
Nozzle at 100-psi Pressure 84
34 Spray Dropsize Distribution for a Delavan 0.75-80°-C
Nozzle Operating at 100 psi 85
35 Spray Dropsize Distribution for a Delavan 1.00-80°-C
Nozzle at 100-psi Pressure 86
36 Spray Dropsize Distribution for a Monarch 1.50-80°-R
Nozzle at 100-psi Pressure 87
37 Spray Dropsize Distribution for a Delavan 0.50-80°-C
Nozzle at 100-psi Pressure 88
38 Spray Dropsize Distribution for a Delavan 0.75-80°-C
Nozzle at 75-psi Pressure 89
vn
-------
FIGURES (Continued)
No. Page
40 Air Flow Parameters at 1.5 Inches Downstream in the
Horizontal Plane of a 55-J Burner 92
41 Air Flow Parameters at 1.5 Inches Downstream in the
Vertical Plane of a 55-J Burner 93
42 Air Flow Parameters at 3 Inches Downstream in the
Horizontal Plane of a 55-J Burner 95
43 Air Flow Parameters at 3 Inches Downstream in the
Vertical Plane of a 55-J Burner 96
44 Air Flow Parameters at 6 Inches Downstream in the
Horizontal Plane of a 55-J Burner 97
45 Air Flow Parameters at 6 Inches Downstream in the
Vertical Plane of a 55-J Burner 98
46 Measured Cold-Flow Air Velocity Vectors in the
Horizontal Plane of a 55-J Burner 99
47 Measured Cold-Flow Air Velocity Vectors in the
Vertical Plane of a 55-J Burner 100
48 Air Velocity Vectors for the Horizontal Plane of
a 55-J Burner 101
49 Air Velocity Vectors for the Vertical Plane of
a 55-J Burner 102
50 Air Velocity Vectors and Fuel Spray Mass Flux Data
for a 55-J Burner Incorporating a 0.75-80°-C Nozzle
(Horizontal Plane) 104
51 Air Velocity Vectors and Fuel Spray Mass Flux Data
for a 55-J Burner Incorporating 0.75-80°-C Nozzle
(Vertical Plane) 105
52 Air Velocity Vectors and Fuel Spray Mass Flux Data
for a 55-J Burner Incorporating a 1.50-80°-C Nozzle
(Vertical Plane) 106
53 Combustion Gas Velocity Vectors in the Horizontal Plane
for the ABC 55-J Burner Mounted in an 8-Inch-Diameter,
Cylindrical, Coaxial Chamber 107
viii
-------
FIGURES (Continued)
No. Page
54 Combustion Gas Velocity Vectors in the Vertical Plane
for the ABC 55-J Burner Mounted in an 8-Inch-Diameter,
Cylindrical, Coaxial Chamber 108
55 Local Combustion Gas Analysis Profiles for the ABC 55-J
Burner at a Nominal Stoichiometric Ratio of 1.25 With a
0.75-70°A Nozzle (Horizontal Plane) 109
56 Local Combustion Gas Analysis Profiles for the BAG 55-J
Burner at a Nominal Stoichiometric Ratio of 1.50 With a
0.75-70°A Nozzle (Horizontal Plane) 110
57 Carbon Monoxide Emissions Measured by Mixed Combustion
Gas Sampling in the Coaxial, 8-Inch-Diameter Cylindrical
Combustion Chamber 112
58 Nitric Oxide Emissions Measured by Mixed Combustion Gas
Sampling in the Coaxial, 8-Inch-Diameter Combustion Chamber 113
59 Smoke Emissions Measured by Mixed Combustion Gas Sampling,
Coaxial 8-Inch-Diameter Combustion Chamber 114
60 Carbon Monoxide Emissions Measured by Lennox Furnace
Flue Gas Sampling 115
61 Nitric Oxide Concentrations Measured in the Lennox Furnace
Flue Gas Samples 116
62 Total Hydrocarbon Emissions Measured by Lennox Furnace
Flue Gas Sampling 117
63 Smoke Emissions Measured by Lennox Furnace Flue
Gas Sampling 118
64 Air Flow Parameters at 1.5 Inches Downstream of the
Mite Burner in an 8-Inch-Diameter, Cylindrical, Coaxial
Chamber 120
65 Air Flow Parameters at 3 Inches Downstream of the Mite
Burner Installed in an 8-Inch-Diameter Cylindrical,
Coaxial Chamber 121
66 Air Flow Parameters at 6 Inches Downstream of the Mite
Burner in an 8-Inch-Diameter Cylindrical, Coaxial Chamber 122
IX
-------
FIGURES (Continued)
No. Page
67 Air Flow and Oil Spray Flux Patterns in the Horizontal
Plane of the Mite Burner Mounted in an 8-Inch-Diameter
Cylindrical, Coaxial Chamber 123
68 Air Flow and Oil Spray Flux Patterns in the Vertical
Plane of the Mite Burner, Mounted in an 8-Inch-Diameter,
Cylindrical, Coaxial Chamber 124
69 Combustion Gas Velocity Vectors in the Horizontal Plane
for the ABC Mite Burner Mounted in an 8-Inch-Diameter,
Cylindrical, Coaxial Chamber 126
70 Combustion Gas Velocity Vectors in the Vertical Plane
for the ABC Mite Burner Mounted in an 8-Inch-Diameter,
Cylindrical, Coaxial Chamber 127
71 Combustion Gas Analysis Profiles for the ABC Mite Burner
at a Nominal Stoichiometric Ratio of 1.25 With a
0.75-70°-A Nozzle 128
72 Combustion Gas Analysis Profiles for the ABC Mite Burner
at a Nominal Stoiochiometric Ratio of 1.50 With a
0.75-70°-A Nozzle 129
73 Combustion Gas Analysis Profiles for the ABC Mite Burner
at a Nominal Stoichiometric Ratio of 1.80 With a
0.75-70°-A Nozzle 130
74 Air Flow and Oil Spray Flux Patterns in the Horizontal
Plane of the Model AFC Burner Mounted in an 8-Inch-
Diameter Cylindrical, Coaxial Combustion Chamber 133
75 Air Flow and Oil Spray Flux Patterns in the Vertical
Plane of the Model AFC Burner Mounted in an 8-Inch-
Diameter, Cylindrical, Coaxial Chamber 134
76 Combustion Gas Analysis Profiles for the Model AFC Burner
at a Nominal Stoichiometric Ratio of 1.25 With a
0.75-70°-A Nozzle 136
77 Combustion Gas Analysis Profiles for the Model AFC Burner
at a Nominal Stoichiometric Ratio of 1.50 With a
0.75-70°-A Nozzle 137
-------
FIGURES (Continued)
No. Page
78 Air Flow and Oil Spray Flux Patterns in the Horizontal
Plane of the Nu-Way Burner Mounted in an 11-Inch-Diameter,
Cylindrical, Coaxial Chamber 139
79 Air Flow and Oil Spray Flux Patterns in the Vertical Plane
of the Nu-Way Burner Mounted in an 11-Inch-Diameter,
Cylindrical, Coaxial Chamber 140
80 Combustion Gas Velocity Vectors in the Horizontal Plane
for the Nu-Way Burner Mounted in an 11-Inch-Diameter,
Cylindrical, Coaxial Chamber 142
81 Combustion Gas Velocity Vectors in the Vertical Plane for
the Nu-Way Burner Mounted in an 11-Inch-Diameter,
Cylindrical, Coaxial Chamber 143
82 Effect of Combustion Chamber Size on Nitric Oxide Emissions 144
83 Smoke Content of the Exhaust Gases for the 6-gph Nu-Way
Burner Operating in the 30-Inch-Diameter, Coaxial Combustion
Chamber Under Cold- and Hot-Wall Conditions 145
84 Hot-Wall, Local Combustion Gas Analysis Profiles for the
6-gph Nu-Way Burner at Nominal Stoichiometric Ratio of 1.16,
With Twin 3-60°-B Oil Nozzles 147
85 Hot-Wall, Local Combustion Gas Analysis Profiles for the
6-gph Nu-Way Burner at a Nominal Stoichiometric Ratio of
1.16, With Twin 3-60°-B Nozzles 148
86 Comparison Plots of Local Combustion Zone Sampled
Stoichiometric Ratio 149
87 Comparison of Combustion Gas Composition Profiles of the
6-gph Nu-Way Model CO Burner in Hot- and Cold-Wall
Enclosures of Different Diameters 152
88 Radiant Energy Profiles for the 6-gph Nu-Way Burner at a
Nominal Stoichiometric Ratio of 1.25 153
89 Radiant Energy Profiles for the 6-gph Nu-Way Burner at a
Nominal Stoichiometric Ratio of 1.50 154
XI
-------
FIGURES (Continued)
No.
90 Diagram of the Carl in Company FFD Flame Funnel Series
Burner Showing the Serai-Staged Combustion 158
91 Nitric Oxide Emissions Obtained From the Carlin 250 FFD
Flame Funnel, the Nu-Way Model CO, and the Sun-Ray Model
PHC Burners 160
92 Air Flow Parameters in the Horizontal Plane at 1.5 Inches
Downstream of the Sun-Ray Burner With No Chamber and
No Oil Flow 162
93 Air Flow Parameters in the Horizontal Plane 3 Inches
Downstream of the Sun-Ray Burner With No Chamber and
No Oil Flow 163
94 Air Flow Parameters in the Horizontal Plane at 6 Inches
Downstream of the Sun-Ray Burner With No Chamber and
No Oil Flow 164
95 Air Velocity Vectors in the Horizontal Plane for the
Sun-Ray PHC-34 Burner With No Chamber and No Oil Flow 165
96 Air Velocity Vectors in the Vertical Plane for the
Sun-Ray PHC-34 Burner With No Chamber and No Oil Flow 166
97 Comparison of Nitric Oxide Emissions in the Flue Gas
for the Sun-Ray Model PHC Burner Operating at 11.5 gph 167
98 Smoke Content of the Flue Gases for the Sun-Ray Model PHC
Burner Fired at 11.5 gph in the 30-Inch-Diameter Coaxial
Combustion Chamber Under Cold- and Hot-Wall Conditions 168
99 Hot-Wall, Local Combustion Gas Analysis Profile for the
Sun-Ray Burner at 11.5 gph and a Nominal Stoichiometric
Ratio of 1.03 169
100 Hot-Wall, Local Combustion Gas Analysis Profiles for the
Sun-Ray Burner at 11.5 gph and a Nominal Stoichiometric
Ratio of 1.03 170
101 Hot-Wall, Local Combustion Gas Analysis Profiles for the
Sun-Ray Burner at 11.5 gph and a Nominal Stoichiometric
Ratio of 1.03 171
XI1
-------
FIGURES (Continued)
•No.
102 Comparison Plots of Local Combustion Zone Sampled
Stoichiometric Ratio 172
103 Nitric Oxide Content of Locally Sampled Combustion Gases,
as a Function of Combusted Fuel Content, for the ABC
55-J Burner 184
104 Nitric Oxide Content of Flue Gases Produced by the
Well-Stirred Burner 187
105 Overall Photograph of the 3- to 12-gph Versatile
Research Burner 190
106 Photographs of the Variable Geometry Burner Heads of
Both the 1- to 3-gph and the 6- to 12-gph Versatile Burners 192
107 Versatile Burner Flue Gas Emissions Obtained With 10-percent
Excess Air in the 30-Inch-Diameter Hot-Wall Chamber 194
108 Versatile Burner Flue Gas Emission Obtained With 10-Percent
Excess Air in the 30-Inch-Diameter Hot-Wall Chamber 195
109 Versatile Burner Flue Gas Emissions Obtained With 10-Percent
Excess Air in the 30-Inch-Diameter Hot-Wall Chamber 196
110 Versatile Burner Flue Gas Emissions Obtained With 10-Percent
Excess Air in the 30-Inch-Diameter Hot-Wall Chamber 197
111 Versatile Burner Flue Gas Emissions Obtained With 10-Percent
Excess Air in the 30-Inch-Diameter Hot-Wall Chamber 199
112 Comparison of Nitric Oxide Emission Results From the 1- to
3-gph Versatile Burner Experiments as a Function of Air
Swirl Vane Angle 201
113 Nitric Oxide Emissions as a Function of Choke Diameter for
the Versatile Research Oil Burner at a Nominal Oil Flowrate
of 0.75 gph with 25-Percent Excess Air 202
114 Choke Diameter Versus Oil Flowrate for Minimum Nitric
Oxide Emissions, Low Smoke, and Carbon Monoxide Emissions 203
115 Nitric Oxide Emission Results From 1-gph Versatile Burner
Experiments in a Perpendicular Port Combustor 206
116 Closeup Photograph of the 1-gph Optimum Oil Burner Head 208
xiii
-------
FIGURES [Continued)
No. Page
117 Nitric Oxide Emissions From the Fixed Geometry, 1-gph
Optimum Burner Fired in the 8-Inch-Diameter, Hot-Wall,
Coaxial Combustor 209
118 Comparison of Nitric Oxide Emission Profiles for Several
1-gph Oil Burners Fired in the 8-Inch-Diameter, Hot-Wall,
Perpendicular Combustor 210
119 Exhaust Gas Composition Profiles as a Function of Time
for the 1-gph Optimum Oil Burner in a Refractory-Lined
Chamber, Hydronic Furnace 212
120 Comparison of Nitric Oxide Emission Profiles for Various
Oil Burners in the 6- to 12-gph Firing Rate Range 214
121 Oscilloscope Frequency Trace of Noise Generated by the
9-gph Fixed Geometry Optimum Burner, Measured at the
Exhaust of the 30-Inch-Diameter, 10-Foot-Long
Coaxial Combustor 216
122 Exhaust Gas Composition Profiles as a Function of Time
for the 9-gph Optimum Oil Burner in a Cold-Wall,
30-Inch-Diameter, Coaxial Combustor 217
123 Nitric Oxide Emissions, Comparison of the Effect of Swirl
on the 30-Degree Convergence, 1.65-Inch-Diameter,
Mechanically Rotated, Six-Vane Swirler, With No Spark Igniter 220
124 Effects of Varying Recirculated Combustion Gas With the
Forced Combustion-Gas Recirculation Burner 223
125 Effects of Combustion Chamber Wall Temperature With
Forced Combustion-Gas Recirculation Burner 225
126 Forced Combustion Gas Recirculation Burner Emissions at
an Oil Flowrate of 0.75 gph 227
127 Schematic of the Two-Stage, Intense-Swirl, Concentric-
Tube Oil Burner Head Extension 229
128 Nitric Oxide Emission Results of Fuel-Rich Combustion
Experiments With a Fixed-Vane Intense-Swirl Burner Head 231
xiv
-------
FIGURES (Concluded)
No._ Page
129 Cross-Section View of the Fixed Geometry, 1.65-Inch-
Diameter No-Swirl Optimum Burner Head 232
130 Nitric Oxide Emissions of the Fixed Geometry, Convergent
and Expansion Sections Head, Optimum Geometry,
No-Swirl Burner 233
131 Variation of the Nitric Oxide Profiles for the 1.65-Inch-
Diameter, Fixed Geometry Versatile Burner With a
0.75-90°-A Nozzle Tested on Different Days 235
132 Comparison of Nitric Oxide Emissions of Commercially
Available Burners With Spark Igniter Modifications 236
133 Comparison of Nitric Oxide Emissions From the 1.65-Inch-
Diameter, No-Swirl, Fixed Geometry, Versatile Burner
With Spark Igniter Modifications 237
xv
-------
TABLES
No... Page
•^^^^^K
1 Contributors to Air Pollution in the United States 16
2 Exhaust Analysis Instruments 55
3 Pure Air Composition 56
4 Validity of Stoichiometric Ratio Calculations of Eq. 7 59
5 Equilibrium Flame Properties for No. 2 Distillate Fuel Oil 60
6 Calculated Equilibrium Gas Composition, Mole Percent 61
7 Commercial Burners Evaluated Experimentally 90
xvi
-------
SUMMARY
This report describes the results of a program to study pollutant forma-
tion and emissions from No. 2 distillate fuel oil burners.
Emphasis was placed on high-pressure atomizing, luminous flame, No. 2
distillate fuel oil burners fired into refractory combustion chambers.
The studies included burners ranging in oil flowrate from 0.5 to 12
gallons per hour (gph).
The objective of this study was to develop information that would allow
minimization of the pollutant emissions from a burner by optimizing the
burner geometry design within practical constraints. The pollutants to
be minimized were nitric oxide, carbon monoxide, smoke, and unburned gas-
eous hydrocarbons.
The approach taken to achieving this objective was to: (1) relate emitted
pollutant concentrations to mixing and flow characteristics of the burner,
(2) relate the mixing and flow characteristics to burner design parameters,
and (3) select the optimum design which minimized pollutant emissions
while maintaining high combustion efficiency. This approach was imple-
mented by: (1) studying the characteristics of several commercially avail-
able oil burners, namely, atomization, oil spray, and air flow patterns in
the combustion zone, combustion gas composition patterns in the combustion
zone, and overall pollutant emissions; (2) interpreting the results from
the commercial burners to determine which design parameters are most im-
portant to the formation of pollutants; (3) building and testing "versatile"
burners capable of continuous variation of the design parameters found to
-------
be most important to pollutant formation; and (4) building and testing
fixed design burners corresponding to the design found from the versatile
burner tests to be optimum for pollutant minimization.
ATOMIZATION MEASUREMENTS
Atomization measurements were made with a number of high-pressure oil
burner nozzles rated for oil flowrates of 0.75 to 1.50 gph. These tests
were conducted with pressure drops from 75 to 125 psig. The frozen wax
method was used for these measurements. Mass median droplet diameters
obtained from these tests ranged from 65 to 90 microns. These results
show that the commercially used atomization nozzles which are available
at very low cost (less than about $1.00 per nozzle, nominally) do an ex-
cellent job of atomization.
TESTS OF COMMERCIAL BURNERS
The air flow patterns produced by several commercially available burners
under noncombustion conditions were determined through the use of a direc-
tional impact probe that could be located at various positions in the oil
burner flowfield. Additionally, oil spray flux patterns were measured with
an aspirated sampling probe. The commercial burners tested in this manner
were the ABC Model 55-J-l (0.75 gph), the ABC Mite (1.0 gph), the Union
Model AFC (0.75 gph), the Nu-Way Model CO (6 gph), and the Sun-Ray Model
PHC (12 gph). Maximum air velocities were found to be, typically, about
30 ft/sec from the 0.75- and 1-gph burners, and about 60 ft/sec from the
6- and 12-gph burners. The measurements of flow patterns were repeated
under combustion conditions with a cooled, directional impact probe being
used. The flow patterns from these tests were found to differ very little,
except for velocity magnitude, from those obtained from cold-flow
measurements.
The same commercial burners were test fired in specially constructed, re-
fractory-lined, cylindrical combustion chambers with the burner blast
-------
tubes mounted coaxially. During these test firings, combustion gas sam-
ples from various locations (including those where velocity measurements
were made) in the combustion zones just downstream of the burner blast
tubes were withdrawn through water-cooled sample probes. The samples were
analyzed for nitric oxide, carbon monoxide, carbon dioxide, unburned hy-
drocarbon, and oxygen content of the sampled gases. The most important
observation from these gas sampling studies was that the amount of nitric
oxide formed, per unit mass of fuel burned, tended to be highest in the
regions where both: (1) the combustion gas composition was near stoichio-
metric, and (2) the velocity patterns tended to promote gas mixing. In
these regions, the nitric oxide concentration tended to be exceptionally
high. The velocity patterns referred to in (2) above included regions
where local combustion gas recirculation was taking place, but away from
the cooling effects of walls and, also, regions where the velocity grad-
ients were relatively steep, thus promoting turbulent mixing. A careful
examination of the experimental data revealed that thorough mixing of the
gas phase, so that the combustion process resembled a well-stirred reactor
(at least locally), led to high concentrations of nitric oxide only when
the local burned gas composition corresponded to 0 to 80 percent excess
air. At local excess air levels greater than 80 percent, the opposite was
true (i.e., thorough mixing led to low nitric oxide emissions). Because
practical oil burners must operate at near stoichiometric conditions (nom-
inally, between zero and about 25 percent excess air) to achieve satisfac-
tory system efficiency, it was concluded that local recirculation, agita-
tion of the gas flow, and steep gradients must be avoided if nitric oxide
emissions are to be minimized. The flame-retention devices, used in the
end of the blast tube of some burners, were found to promote the unwanted
local recirculation, agitation, and steep velocity gradients. Therefore,
the flame-retention device was found to be undesirable from a pollution
standpoint.
-------
VERSATILE BURNER EXPERIMENTS
Based on the results from the commercial burner studies, versatile geom-
etry burners were constructed that simulated the construction of nonflame-
retention burners. The variable geometry features of these burners in-
cluded a continuously variable choke diameter, variable nozzle recess,
and six peripheral, variable-angle air swirler vanes. Two such versatile
geometry burners were constructed, one for use with flowrates of 1- to
3-gph (4-inch-diameter, 9-inch-long blast tube), and one for use with
flowrates of 3- to 12-gph (5-inch-diameter, 12-inch-long blast tube).
Combustion experiments were conducted with these versatile burners in
refractory-lined, coaxial, cylindrical combustion chambers, over the 1-
to 12-gph range of oil flowrates and over practical ranges of choke diam-
eters, air swirler vane angles, and oil spray nozzle types. During all of
these experiments, the pollutant emissions were measured. The results
showed that usually the smoke, unburned hydrocarbons, and carbon monoxide
emissions decreased, often monotonically, with increasing air swirler vane
angle, whereas nitric oxide emissions increased, often monotonically,
also with increasing air swirler vane angle. An air swirler vane angle of
25 degrees (measured relative to the blast tube axis) was found to be a
good compromise over the entire oil flowrate range from 1- to 12-gph. The
primary effect of the blast tube choke diameter was on the nitric oxide
emissions, the choke diameter for minimization of nitric oxide emissions
being given by choke diameter (inches) = (2.7 x gph)0-4. The oil spray
atomizer characteristics and oil nozzle recess were found to have little
effect on the emissions, with the exception that nozzles producing smaller
dropsizes (i.e., small orifice nozzles operated at high pressure drop)
tended to produce less smoke than the larger dropsize nozzles (i.e.,
large orifice nozzles operated at low pressure drop). Oil nozzle position
and spray angle had little effect on emissions in the versatile geometry
burner studies, although later studies with fixed geometry burners re-
quired a trial and error selection of oil nozzle spray angle necessary for
elimination of smoke emissions, with the best spray angle apparently being
dependent on chamber geometry (implying that a nozzle must be selected
-------
which eliminates impingement of raw oil on the chamber walls). The opti-
mum burner geometry criteria, therefore, reduced to utilizing six peri-
pheral air swirler vanes at an angle of 25 degrees relative to the blast
tube axis, and a choke diameter given by (2.7 x gph)°-4.
OPTIMIZED FIXED GEOMETRY BURNERS
Fixed geometry burners, sized to burn nominally 1-gph and 9-gph of No. 2
fuel oil, were designed according to the optimum criteria obtained from
the versatile geometry burner results. Combustion tests were conducted
with the burners to verify the low pollutant operation.
1-gph Optimum Burner
The 1-gph optimum, fixed geometry burner was test fired in a refractory
lined, coaxial combustion chamber. As expected from the versatile burner
geometry optimization experiments, the optimum burner produced negligible
smoke, unburned hydrocarbon and carbon monoxide emissions, and the nitric
oxide emissions were only ^1 gm NO/kg fuel burned, which may be compared
with nominal levels of 1.5 to 2 gm NO/kg fuel from typical commerical
burners fired into the same combustion chamber. Additionally, the optimum
geometry burner was capable of operating at only 5 to 10 percent excess
air without smoke, whereas typical commercial burners require up to 25
percent excess air to achieve smoke free operation.
Moreover, when the 1-gph optimum, fixed geometry burner was fired into an
air-cooled combustion chamber, the nitric oxide emissions were reduced by
20 percent relative to similar emissions obtained in a hot-wall chamber.
When the 1-gph optimum, fixed geometry burner was fired into a refractory
wall, perpendicular port combustion chamber, with the combustion chamber
axis perpendicular to the axis of the burner blast tube, the nitric oxide
emissions increased to a nominal 2 gm NO/kg fuel, compared with nitric
oxide emissions of 2 to 3 gm/kg for typical commerical burners fired into
-------
the same chamber. The higher emissions in the perpendicular port combus-
tion chamber apparently resulted from a high level of gas-phase mixing
caused by the right-angle turn from the burner blast tube axis to the
combustion chamber axis. As noted above, strong mixing in the combustion
zone tends to induce high nitric oxide emissions.
The 1-gph optimum fixed geometry burner was subjected to simulated field
tests, accumulating a total time of 128 hours while being cycled on and
off to simulate field usew The tests were conducted in a Unitron A100
hydronic/warm air furnace having a coaxial, refractory-lined combustion
chamber. During this testing, no adjustments were made on the burner.
The 1-gph optimum burner performed over the entire test period with little
variation and no noticeable degradation. The nominal stoichiometric ratio
drifted upward slightly (about 2 percent, from 1.08 to 1.10 to 1.10 to
1.12), probably due to changes in ambient conditions associated with the
outdoor test facility. None of the exhaust pollutant emission concentra-
tions increased during the testing, with nitric oxide at about 0.95 gm
NO/kg fuel burned, no measurable unburned gaseous hydrocarbons, and no
smoke (Bacharach smoke scale = 0). Carbon monoxide decreased slightly
(0.60 ->• 0.35 gm CO/kg fuel), probably due to the shift in stoichiometric
ratio. Removal and inspection of the 1-gph optimum burner after comple-
tion of testing revealed no soot or sludge accumulation nor any evidence
of deterioration of either the burner or the furnace.
9-gph Optimum Burner
The 9-gph, optimum, fixed geometry burner was test fired only in a co-
axial, water-cooled combustion chamber. The inside diameter of the com-
bustion chamber was 30 inches, nominally sized for 12 gph. The nitric
oxide emissions from the 9-gph burner were very low, at a level of about
0.6 gm NO/kg fuel, compared to 1.2 gm NO/kg fuel for both the 6- and 12-
gph commercial burners when fired into the same combustion chamber. The
9-gph optimum burner was found to be capable of operating smoke free at
as low as 2-percent excess air, whereas the commercial burners required
from 5- to 25-percent excess air to eliminate smoke.
-------
Simulated field testing of the 9-gph optimum burner also was conducted,
with a total of 112 hours being accumulated with this burner in the 30-
inch-diameter, water-cooled, coaxial combustor. The performance of this
larger optimum burner also showed little variation and no noticeable
degradation. The stoichiometric ratio increased slightly from 1.02 to
1.03 to 1.03 to 1.04. A slight improvement was noted in the NO emission
level, which decreased from about 0.5 to 0.6 gm NO/kg fuel to 0.45 to
0.50 gm NO/kg fuel burned. The CO emission level varied slightly, prob-
ably with changes in ambient conditions, but emissions as low as 0.6 to
0.7 gm CO/kg fuel burned were still being obtained at the end of the test
series. The emission level of unburned hydrocarbons was zero and the
Bacharach smoke number was zero. Inspection of the 9-gph optimum burner
at the end of the test series revealed only a slight oily accumulation
around the perimeter of the 5-inch-diameter burner head, and no evidence
of deterioration of the burner head assembly.
The overall results from the optimum fixed geometry burner tests are con-
sidered very good. They show that the design criteria generated from the
versatile geometry burner experiments do lead to low pollution burner de-
signs. The optimum geometry burners are capable of operation at very low
excess air levels while still producing emissions significantly lower than
presently obtained with commercial burners. Low excess air levels lead to
high furnace efficiency because of the resultant high flame temperatures
that benefit heat transfer rates.
NONCONVENTIONAL BURNER CONCEPTS
In addition to the above-described studies of conventional, high-pressure
atomizing, luminuous-flame burners designed for operation in refractory-
lined combustion chambers, a portion of the program effort also was di-
rected toward nonconventional burners. Several types of nonconventional
burners were investigated, with the most notable results being obtained
with an intense swirl burner and a forced combustion gas recirculation
burner.
-------
The intense swirl burner contained a mechanically rotated paddle wheel
located at the end of the burner blast tube for swirling the inlet combus-
tion air. The paddle was rotated by an electrical motor at rates as high
as 3450 rpm. Combustion gas analyses from samples obtained in the com-
bustion zones of commerically available burners had led to the conclusion
that such an intense swirl burner would produce very high nitric oxide
emissions at excess air levels near stoichiometric, but very low nitric
oxide emissions at extremely high excess air levels (in excess of 80-
percent excess air). The intense swirl burner behaved essentially as pre-
dicted from the commercial burner data. The nitric oxide emissions were
very high at low excess air levels (about 3 gm NO/kg fuel burned at 25-
percent excess air) and very low at high excess air (about 0.35 gm/kg
fuel at 120-percent excess air). The smoke, unburned hydrocarbon, and
carbon monoxide emissions were all minimal for these tests. The tests
were conducted in a perpendicular port, refractory-lined combustion cham-
ber, and the low nitric oxide emissions observed at high excess air levels
were lower than those observed in commercial burners (operated in similar
combustion chambers) by a factor of from 3:1 to 5:1. Operation at such
high excess air levels requires a very large heat exchanger to maintain the
furnace efficiency at a high level, because the low flame temperatures
coincident with the high excess air levels reduce heat transfer efficiency.
Operation at these high excess air levels is, therefore, unacceptable, and
the intense swirl burner was regarded as a tool to investigate combustion
phenomena rather than as a commercially viable burner.
The forced combustion-gas recirculation burner was designed for operation
in either air-cooled or water-cooled combustion chambers. In this burner,
a fan is used to draw combustion gases out of the combustion chamber from
the zones near the wall where the gases are likely to have been cooled by
convective heat transfer to the wall. These combustion gases are mixed
with the burner primary air and reinjected into the chamber. Thus, this
burner was designed to achieve the same benefits as those obtained with
flue gas recirculation without requiring a duct from the flue to the
burner.
-------
When this burner was tested, the nitric oxide emissions were found to be
very low, about 0.3 to 0.4 gm NO/kg fuel, at nearly stoichiometric condi-
tions, with no measurable smoke, and very low emissions of carbon monoxide
and unburned hydrocarbon. In some cases, this burner was found to oper-
ate under fuel-rich conditions with no significant smoke. When this
burner was tested in refractory wall combustion chambers, significantly
increased nitric oxide emissions were obtained, as expected, becuase of
the higher temperature of the recirculated combustion gases. Increasing
the oil flowrate of this burner from 0.5 to 0.75 gph, in a cooled chamber,
also resulted in increased nitric oxide emissions, apparently due to a
limitation in the available amount of cooled combustion gases near the
walls of the combustion chamber. The forced combustion gas recirculation
burner, or some modification thereof, appears to be a good choice for the
burner design when a nonrefractory-lined, cooled-wall combustion chamber
is available.
CONCLUSIONS
The major conclusions from this study may be summarized as follows:
1. In refractory-lined combustion chambers, minimum nitric oxide
emissions are obtained from nonflame-retention burners having
peripheral swirlers at 25-degree angles (relative to the burner
blast tube axis) and choke diameters corresponding to:
0 4
Dia. (inches) = (2.7 x gph) ' .
2. Variations of oil nozzle position (0- to 1-inch recess), oil
pressure drop (75 to 125 psi), and spray cone angle (45 to 90
degrees) affected smoke emissions but had no significant
effects on other pollutants. Proper matching of the atomizer
spray angle to the combustion chamber geometry is necessary to
avoid smoke emissions.
3. Coaxial combustion chambers result in lower pollutant emissions
than perpendicular port combustion chambers. This has been
demonstrated for refractory lined combustion chambers.
-------
4. Cool-wall combustion chambers result in lower nitric oxide emis-
sions than refractory wall combustion chambers.
5. In cool-wall chambers some form of combustion gas recirculation
is beneficial to the minimization of nitric oxide emissions,
but in refractory wall combustion chambers, this recirculation
is not as beneficial.
SUMMARY OF RECOMMENDED DESIGN PRACTICES
Design recommendations for minimizing exhaust emissions have been devel-
oped especially for high-pressure atomizing, gun-type oil burners firing
into adiabatic (uncooled) wall combustion chambers, with emphasis being
placed on remaining within the current conventional oil bu rner manufac-
turing practices. These optimized burner criteria are also applicable to
cooled-wall combustion chambers, but under cooled-wall conditions, other
methods (i.e., "blue flame" burners and combustion/flue gas recirculation
burners) are available to reduce pollutant emissions further.
Combustion Chambers
The coaxial combustion chamber configuration was found to be more favorable
for obtaining low pollutant emissions than the perpendicular burner port
enclosures. Design of combustion chambers with the burner insertion port
aligned along the axis of the chamber is recommended. A cooled chamber
wall also is favorable in achieving lower nitric oxide emissions levels.
Air Swirl Devices
Omission of flame-retention devices (i.e., flame cones and flame-retention
swirler rings) is recommended. However, some air swirl (rotational energy)
is needed to promote mixing of the air and the oil spray. Simple peri-
pheral swirl vanes (six vanes) that extend from the blast tube perimeter
to within 0.3 inch of the oil nozzle assembly, and having a length/width
10
-------
ratio of about 1.5, are recommended. Vanes that project radially, that
are spaced equally around the burner head, and are set at an angle of 25
degrees relative to the blast tube axis (Fig. 1) are recommended. Minor
modifications to the existing spark electrode designs will probably be
required to avoid interference with the large vanes. The configuration
shown in Fig. 1 was typical of those tested during this study. Other
slightly different configurations, designed for greater ease of fabrica-
tion or servicing, may be equally acceptable.
Choke Diameter
The choke diameter (ex.it area) has a significant effect on nitric oxide
emissions. Minimum nitric oxide is obtained with a diameter given by the
following relationship:
Dia = (2.7 x gph)°'4
where
Dia = optimum choke diameter, inches
gph = oil flowrate, gal/hr
This correlation is shown graphically in Fig. 2, where the shaded area
represents the suggested allowable deviation from the specified diameter.
Oil Nozzles
The high-pressure (100 to 300 psig), swirl-atomizing, simplex nozzles are
entirely adequate oil atomizing devices. In larger oil burners, the use
of higher oil supply pressure (300 psig) or multiple nozzle assemblies
(which are frequently used in current commercial practice) tends to reduce
smoke emissions because smaller oil dropsizes are formed. The additional
pump power requirements of the higher oil supply pressure system suggest
the multiple nozzle system is the more favorable of these two design
recommendations.
11
-------
(a) Closeup of a 1-gph optimized oil burner head
(b) Optimized burner head mounted on a
typical oil burner body
Figure 1. Photographs of a 1-gph, low pollutant emission
optimized oil burner head
12
-------
10.0
8.0
1.0
0.6
0.8 1.0
2.0 4.0
Oil Flowrate, Gallons/Hour
6.0
i.O 10.0
20.0
Figure 2. Choke diameter versus oil flowrate for minimum nitric
oxide emissions, low smoke and carbon monoxide emissions
(Applicable to burners incorporating a six-vane, 25-degree
air swirler)
-------
Electrical Spark Igniter
The operation of the 10,000- to 12,000-volt electrical spark igniter has
shown some significant effects on nitric oxide emission levels from non-
flame-retention-type burners. Avoidance of continuous operation of the
spark igniter throughout the burner firing time is recommended, although
continuous spark may be necessary to ensure nonoscillatory combustion.
Burner designs that incorporate spark operation only during burner start-
up are recommended. Considering the power consumption of a typical spark
igniter of 250 watts (MOOO Btu/hr), this recommendation should reduce
the power requirement significantly.
14
-------
INTRODUCTION
The combustion of fossil fuels currently generates approximately three-
fourths of all air pollution. As shown in Table 1, the major contribu-
tors to overall air pollution are: 50 percent from transportation (cars,
trucks, and airplanes), 20 percent from stationary fuel combustion (elec-
trical power generation and space heating), and 13 percent from industrial
processes. Stationary fuel combustion contributes 28 percent of the parti-
culates and 39 percent of the nitric oxide emissions.
Of particular interest to this program is the air pollution caused by dis-
tillate oil burners for commercial or residential space heating. During
the peak heating season, such oil burners contribute heavily to the parti-
culate and nitric oxide pollution in urban areas. They also contribute un-
burned hydrocarbons and carbon monoxide to a less significant extent.
In large measure the control and/or elimination of such emissions is di-
rectly related to the care and maintenance of oil burning equipment. Hy-
drocarbons and particulate matter, for example, can be virtually eliminated
by proper adjustment of fuel oil burners and careful maintenance to avoid
degradation. Burner combustion efficiency and emissions of oxides of ni-
trogen are interrelated parameters which can be varied by burner design
and operating characteristics. Emissions of oxides of sulfur are largely
determined by the fuel and can be restricted through the use of low sulfur
fuel.
15
-------
Table 1. CONTRIBUTORS TO AIR POLLUTION IN THE UNITED STATES
Type of Pollutant
Carbon Monoxide
Sulfur Oxides
Hydrocarbons
Particulates
Nitrogen Oxides
Total
Percent of Total
Trans-
portation
Stationary
Fuel
Combustion
Industrial
Processes
Solid
Waste
Disposal
Quantity, millions of tons per year
71.2
0.4
13.8
1.2
8.0
94.6
50
1.9
22.1
0.7
6.0
6.7
37.4
20
7.8
7.2
3.5
5.9
0.2
24.6
13
4.5
0.1
1.4
1.2
0.7
7.9
4
Miscel-
laneous*
8.6
0.6
6.5
7.2
1.4
24.3
13
Source: National Air Pollution Control Administration
*Forest fires, etc.
The primary pollutants emitted into the atmosphere by oil burners are:
oxides of nitrogen, smoke, carbon monoxide, and gaseous hydrocarbons. All
except nitrogen oxides can be minimized by operating at sufficiently high
air-to-fuel ratios. Unfortunately, operation at high air-to-fuel ratios
generally leads to lowered performance, and frequently to higher nitric
oxide emissions.
The broad purposes of the program described in this report were to deter-
mine the effects of burner design parameters on pollutant emissions for
stationary energy conversion devices that use No. 2 fuel oil, and to de-
fine designs which allow pollutants to be minimized while maintaining
high efficiency. The effort has been directed primarily toward reduction
of pollutant emissions from high-pressure atomizing, 1- to 12-gph No. 2
distillate oil burners fired in refractory combustion chambers.
The typical oil burner of this type, Fig. 3, has spark ignition,
achieves oil atomization by pumping the oil at high pressure (100 to
16
-------
FLANGE
MOUNTING
IGNITION
TRANSFORMER
1
SWIRL VANES/ V!l^>
CAD CEIL BURNES -MOUNTED CONTROL
ELECTRODE LEADS
FUEL TUBE
ELECTRIC MOTOR
HOUSING
Figure 3. Typical high-pressure, oil-fired gun burner
-------
300 psig) through a conical spray simplex nozzle, Fig. 4, and supplies
air for the combustion process through an electrically driven blower.
The oil nozzle and spark ignition electrodes are located in the end of
a cylindrical tube (blast tube) through which the combustion air flows
into the combustion chamber. The design of the air diffuser located at
the blast tube end (Fig. 5) strongly affects the air flow pattern and,
hence, the combustion process. Many different blast tube end designs
are available, ranging from a sinple choke ring (an orifice plate affixed
to the end of the blast tube to restrict the air flow) to elaborate swirl
vane and flame-retention assemblies intended to enhance mixing and to pro-
vide regions of low-velocity air/fuel-spray mixture to allow flame attach-
ment. The levels of pollutant emissions from such commercially available
burners vary significantly, depending on the design selected. Several oil
burner manufacturers were contacted to determine present state-of-the-art
burner design practices. The information obtained from these contacts is
summarized in Appendix A.
This study has been directed toward defining burner modifications that
would minimize emissions and maximize efficiency of the high-pressure gun
burner, and be economically feasible to manufacture, install, and operate.
The economics are important because there are almost 12 million oil burners
in operation in the United States today, of which well over 90 percent are
the high-pressure gun burner type. An example of a desirable device for
pollution control would be one that could be simply and economically in-
stalled at the time of fuel nozzle replacement. This would result in
rapid improvement of the existing oil burner installations because high-
pressure fuel nozzles are normally replaced every 2 to 3 years. In con-
trast, a device requiring major burner modification or replacement of
significant portions of the burner assembly and its auxiliary components
would be less desirable.
Burner design and operation are significantly affected by the type of com-
bustion chamber in which the burner is operated. The largest number of oil
18
-------
DISCHARGE
ORIFICE
SWIRL CAVITY
.SLOTTED SWIRL PLUG
•FLOW DISTRIBUTOR BODY
Figure 4. Simplex oil spray nozzle
FLAME RET:
CONE
ANGULAR VANES
Figure 5. Three typical air diffuser designs used
in the ends of oil burner blast tubes
19
-------
burners are used in furnace installations. Most modern oil furnaces are
warm air delivery systems; however, a large percentage of the heating in-
stallations in the northeast are hydronic (warm water) systems. The com-
bustion chambers in the furnaces vary widely in configuration and char-
acteristics; for example, many of the older warm air furnaces have a heavy
refractory lining and the chambers are large. The most recent warm air
combustion chambers contain a lightweight, cast refractory lining and are
smaller. Many hydronic systems have one or more cold walls exposed to the
combustion system. The predominant number of residential furnace combus-
tion chambers are refractory or refractory lined and, therefore, this
study has been centered primarily on improvement of burners operated in
refractory combustion chambers.
A wide variation in fuel flowrates is used in existing installations. Al-
most all of the residential heating installations are operated in a range
from 0.5 to 1.2 gal/hr, with 80 percent in the 0.6 to 0.9 gal/hr range.
Commercial installations, however, are operated predominantly in the 1 to
3 gal/hr range, with some higher flowrates also used. Some oil-fired,
hot-water heaters operate with flowrates as low as 0.5 gal/hr. Conse-
quently, an air pollution-improving device should be effective over this
wide flowrate range. The major portion of effort in this study has con-
cerned the nominal flowrate of 1 gph, with a significant but lesser effort
concerned with oil flowrates up to 12 gph.
All oil burners must meet stringent safety standards to be acceptable for
installation in the United States. Safety standards are set by the Under-
writers Laboratory (UL) and conformance is evaluated by tests. The basic
UL specification applying to the high-pressure oil burner design is UL
Specification 296, "Standard for Domestic Oil Burners." Some of the UL
requirements related to the burner head are as follows: (1) ability to
remove the ignition electrodes and fuel nozzle assembly without removal
of the entire burner, (2) high-tension wiring must be installed with a
minimum separation from other components, (3) materials having sufficient
temperature capability must be used within the combustor head, and
20
-------
(4) specific provision must be made for oil dripping from the fuel nozzle
and for handling oil that may accumulate on the electrodes. Any burner
design modifications must be compatible with these requirements.
The role of the serviceman is paramount in the oil equipment business.
Installation, maintenance, and service training requirements for new equip-
ment must be minimal to provide an acceptable product. Excessive complex-
ity of any device could prevent its use.
Because all oil burners deteriorate with time, a desirable characteristic
of a pollution control device would be insensitivity to degradation. A
survey conducted by Rocketdyne prior to start of the present study indi-
cated that the typical oil furnace installation set for less than No. 1
Bacharach smoke at the initiation of a heating season usually operates at
greater than No. 4 smoke by the end of the heating season. This indicates
the need for more frequent service, at least once per year compared to the
typical once every 2 or 3 years.
In view of the safety, cost, serviceability, furnace compatibility, and
degradation factors, and the basic program objective of producing low pol-
lutant emissions in conjunction with high efficiency, the following ap-
proach was selected for this program: (1) study the mixing, atomization,
and flow characteristics of several commercially available burners; (2)
determine the pollutant formation and emission characteristics of these
burners; (3) relate the mixing, atomization, and flow characteristics to
the pollutant formation and emission; and (4) determine optimum burner
design criteria that would lead to the atomization, mixing, and flow char-
acteristics yielding the best compromise between low pollutant emissions
and high efficiency.
21/22
-------
APPARATUS AND DATA REDUCTION
This section describes the apparatus used to characterize burner processes
experimentally, and, where appropriate, includes the methods used to re-
duce the experimental data to reported form. The apparatus described in-
cludes those items other than the commercially available burners. These
burners are described later, together with data obtained from each.
ATOMIZATION TEST FACILITY
A wax flow technique was used to determine the dropsize distribution pro-
duced by the oil spray nozzles. The facility used for those tests is
shown in Fig. 6. The system consists of two molten-wax tanks, one hot-
water tank, pneumatic control valves, and a thermostatically controlled
oil bath vessel in which the wax and water tanks are immersed. Associated
flow and collection equipment includes Taber injection pressure trans-
ducers, turbine flowmeters, a particle collector surface, and a particle
catch basin into which the particles are washed from the particle collec-
tor. This facility is used extensively at Rocketdyne for characterization
of rocket engine injectors. The hot-water flow system is used to simulate
bipropellant rocket flows, but was not required for the oil burner atomizer
flows.
The hot-oil bath is heated by a 30-kilowatt, thermostatically controlled
heater. An electrically operated pump circulates the oil from the oil
bath container through the heater and back again to ensure uniform temper-
ature. In addition, hot oil is forced through jacketed run lines and
23
-------
5AD21-6/13/69-S1B
(a) Wax flow system
5AD21-6/13/69-S1C
(b) Particle collector platform
Figure 6. Wax flow facility
24
-------
valves to ensure that the wax does not freeze or cool prior to spraying.
The wax flow system has three parallel line sections each containing a
flowmeter, a thermocouple, and a hand shutoff (flow control) valve. A
wide range of flowrates can be obtained by opening one of the three hand
shutoff valves leading to the appropriate flowmeter spanning the desired
flowrate range.
Between the atomizer and the frozen droplet collector surface, a deflector
tube is used that ducts the wax spray away from the collector until steady-
state wax flow conditions are established. Through use of a high-speed
pneumatic actuator, the tube is removed for the duration of the run and is
replaced prior to flow cutoff, thus eliminating wax-particle collection
during flow start and cutoff transients. The particle collector is an 18-
by 50-foot, epoxy-coated, wooden platform that slopes gradually toward the
center of the platform and away from the atomizer. The entire platform is
located under a semi-enclosed structure that shields the collection area
from wind currents that might cause the smaller particles to be blown
away. The slope of the platform causes the wax droplets to be directed
into a relatively small-particle catch basin when the impact surface is
washed down with water. The catch basin has several baffles to ensure
that none of the wax particles are washed overboard.
For the present study, wax flow tests were made with the oil burner ato-
mizer nozzles mounted in the end of a hot-oil-jacketed wax flow line, with
no attempt being made to simulate the air flow patterns produced by typi-
cal oil burners. This was believed valid because the nature of the coni-
cal spray nozzles being used causes most of the atomization to take place
very close to the nozzle exit where the liquid films are shielded from air
flow by the body of the atomizer.
The wax utilized for the spray dropsize experiments was Shell Type 270 wax
(Petrochemicals Division, Shell Chemical Company). This wax has a melting
point of 140 F, and was sprayed at a temperature of approximately 200 F.
The Type 270 wax was chosen because of its high block point temperature
25
-------
of 120 F, which implies that the wax particles are not sticky at tempera-
tures below 120 F. The nonstick/ quality of the wax particles is very
important for particle size determination. The high block point wax al-
lows the collected wax particles to be sieved much like dry sand, whereas
lower block point waxes yield particles which tend to congeal into sticky
masses which are very difficult to sieve. At the 200 F spraying tempera-
ture, the Type 270 wax has a viscosity of 0.0027 Ibm/ft-sec, a density of
48 Ibm/ft^, and a surface tension of 17 dynes/cm. The viscosity and den-
sity of Type 270 wax are reasonably comparable with the viscosity and
density (0.0026 Ibm/ft-sec and 56 lb/ft3) of No. 2 fuel oil.
For the nozzles tested, the mass median droplet diameters, D", of the col-
lected frozen wax particles were in the 60- to 100-micron (25.4 microns =
0.001 inch) range. The mass median droplet diameter is defined as the
diameter for which one-half of the spray mass has larger diameters and the
remaining half has smaller diameters. For the 60- to 100-micron range,
the size distribution of the collected particles was most conveniently
determined by the Sharpies Micromerograph, a commercially available sedi-
mentation- in-air device. Particles from one test run were also sized by
sifting through a series of standard testing sieves. For the test on
which both sizing techniques were utilized, size distribution data ob-
tained by means of the Micromerograph were found to yield a mass median
droplet diameter approximately 10 percent larger than was obtained from the
parallel analysis using the sieving technique. This discrepancy is prob-
ably due to the loss of small particles to the wall of the Micromerograph,
but its magnitude is small.
AIR FLOW PATTERNS
Air flow patterns were determined by measuring both the direction and mag-
nitude of air velocity at many points throughout the flowfield created by
the oil burners.
26
-------
Air flow patterns were determined near the blast tube exits of several
burners, with the flow constrained within short, cylindrical sheet metal
chambers mounted coaxially with the burner blast tubes. For the nominal
1-gph burners, an 8-inch-diameter coaxial cylindrical chamber was used
while, for the 6- and 12-gph burners, an 11-inch-diameter coaxial cylin-
drical chamber was used. The coaxial chamber extended approximately 9
inches beyond the end of the burner blast tube, and air velocity vector
profiles were determined at several planes perpendicular to the blast tube
axes, at distances up to 9 inches downstream from the blast tube exit.
Two different gas velocity-vector direction-indicator probes were built
for initial checkout. These two probes are shown schematically in Fig. 7.
They were designed to make use of the fact that the pressure indicated by
an impact tube varies most rapidly with angular orientation when the tube
inlet is inclined at 45 degrees to the flow. Flow direction in one plane
was determined by changing the probe angle until the indicated pressures
from opposed impact tubes with inlets inclined at 90 degrees to one
another were equal. Flow direction in the other plane was obtained by
similarly balancing the second pair of impact tubes of the quadruple tube
probes. For this purpose, the probe was mounted on a gimbaling apparatus,
as shown in Fig. 8 which allowed changes in probe angle without changing
the location of the probe tip. The probe was mounted so that one of the
adjustable angles corresponded to movement in the plane of one pair of
probe tubes, and the remaining adjustable angle corresponded to the plane
of the other pair of tubes. The two gas velocity-vector direction-indica-
tor probes shown in Fig. 7 differ primarily in the manner by which the
tube inlets are inclined to the axis of the probes. In the compact probe
(Fig. 7a), the inclined inlets result from cutting each of the tubes in
the four-tube bundle at 45 degrees to the bundle axis while, with the
standard probe, the inlet inclination was obtained by bending the tubes as
shown in Fig. 7b. it was expected that the standard probe (Fig. 7b)
would give more precise information concerning the flow angle than the
compact probe; however, the experimental results indicated no significant
differences between the two probe types. Therefore, the compact probe
27
-------
(A) Compact Probe
(Uncooled)
Q
CD QD
(B) Standard Probe
(Uncooled)
Section X-X
Figure 7. Velocity direction indicator probes
28
-------
Lathe Head
Figure 8. Schematic of directional probe gimbaling apparatus
used to alter probe angle while keeping the probe
tip stationary
29
-------
probe design, which offers less interference with the flow and the combus-
tion chamber walls, was used for determination of gas velocity vector di-
rections. The probe was calibrated in gas flows of known direction. Once
calibrated, the probes gave direction indications which were reproducible
to within ±2 degrees, with oil burner air flows of about 10 ft/sec. The
primary factor limiting reproducibility appeared to be nonsteadiness of
the flows produced by the oil burner.
The probes shown in Fig. 7 indicate only the vector direction of the gas
flow. Another probe was used to determine the magnitude of the air velo-
city, a commercially available 1/16-inch-OD pitot probe. After determin-
ing the flow vector directions, the multiple-port impact probe was re-
placed on the gimbaling apparatus by the pitot probe. The pitot probe
was aligned with the previously determined air velocity vector directions
before making a measurement of the velocity head. For both the multiport
vector direction probe and the pitot probe, the required pressures were
measured with an MKS Baratron electronic pressure meter. A 30-millimeter
pressure sensor was used in conjunction with the Baratron which allowed
measurement of impact pressures as low as 0.001 mm Hg.
The apparatus used for these experiments is shown in Fig. 9. The gimbal-
ing device was fabricated from a rotatable lathe, used for one of the
gimbal directions, and a tube and aluminum I-beam structure for the other
gimbal direction. In addition to the effective gimbaling action available
from rotation of the lathe table, the lathe table allowed translational
movement in two directions. Also shown in Fig. 9 is the 55-J burner and
the Baratron electronic pressure meter.
With the gimbaling apparatus shown in Fig. 8, it was possible to define
the flow vector at any point of interest in terms of two angles and an
impact velocity head. The angle of rotation provided by the rotatable
30
-------
55-J BURNER
ROTATABLE
LATHE TABLE
PITOT PROBE
(a) Set for a compound angle
(b) Set for straight axial flow
Figure 9- Probe gimbaling apparatus
31
-------
lathe head was termed a, and the angle of rotation provided by the tube
and I-beam structure was termed 3. With these parameter definitions, the
components of velocity are obtained from the following equations:
VT = V sin 3 . . ,
/2 gc AP \1/2
VR = V cos 3 sin a, V =1 ] (l)
VA = V cos 3 cos a ^ '
where
V = velocity scalar magnitude, ft/sec
VT s tangential velocity component, ft/sec
VR = radial velocity component, ft/sec
VA = axial velocity component, ft/sec
a = velocity angle, measured in degrees of rotation about an
axis perpendicular to both the burner tube axis and the radial
ray being traversed
3 = velocity angle, measured in degrees of rotation about an axis
inclined at an angle a to the radial ray being traversed, and
located in the plane established by the burner tube axis and
the radial ray being traversed
gc = gravitational constant, ft-lbm/lbf-sec^
AP = impact head, lbf/ft2
p = gas (air) density, lbm/ft3
To interpret the data giving the 3 angle as a function of location, it is
helpful to realize that a curve of 3 versus radius consisting of a straight
line passing through the burner axis (r = 0), such as shown in Fig. 10,
corresponds to a flowfield rotating at a uniform rpm about the axis, while
the dashed curve corresponds to a flowfield rotating at a higher rpm on the
outer radii, and the dotted curve corresponds to a flowfield rotating at a
32
-------
Radius, r
Uniformly Rotating
Flowfield
Higher RPM on the
Outer Radii
Higher RPM on the
Inner Radii
Figure 10. Interpretation of 6 angle curves
33
-------
higher rpm at the inner radii. Strictly speaking, these generalizations
require flowfield symmetry which is not obtained with most burners tested,
but the trends are applicable.
Prior to the start of testing for any particular burner, the velocity vec-
tor direction probe was calibrated by flowing gaseous nitrogen through a
3/8-inch-diameter tube and impinging on the probe. Generally, the 3/8-
inch-diameter calibration tube was aligned parallel with the axis of the
burner blast tube. When the gimbal was used to orient the direction of
the multiport impact tube to obtain null pressure readings for both planes,
the a and B angle scales were then adjusted to each read zero degrees,
indicating precisely axial flow. The traversing capabilities of the lathe
table were used to accurately move the probe tip from point to point in
the flowfield. The lathe table allowed translation of the probe tip in a
plane corresponding to a major diameter of the blast tube axis, and in the
axial direction corresponding to various distances downstream of the blast
tube exit. For these movements, the axis of rotation of the a angle was
maintained in the same plane. To make measurements in two perpendicular
planes required mounting the burner twice. In general, measurements were
made in two planes which were termed the horizontal and the vertical
planes, as shown schematically in Fig. 11.
OIL SPRAY PATTERNS
Under cold-flow conditions, oil spray mass fluxes were determined as a
function of location in the gas flowfield. To obtain meaningful oil spray
mass flux data, it is necessary to use an oil spray collector that is
aligned with the local air flow velocity vector. It is also necessary to
sample isokinetically, so that air velocity entering the sample tube is
equal to the local air velocity. If air is entering the sample tube at
less than local air velocity, a high-pressure (partial stagnation) zone
is created near the tip of the sample tube, deflecting droplets away from
the sample tube inlet. On the other hand, if air is pulled into the tube
at a higher than local velocity, a low-pressure zone is created near the
sample tube inlet, and more than the desired oil sample rate is obtained.
34
-------
Vertical
Plane
End View of Burner
Blast Tube Axis
Horizontal Plane
Motor
Bottom
Figure 11. Location of horizontal and vertical planes
for determination of gas velocity vectors
35
-------
Apparatus used for determination of oil spray mass flux under isokinetic
sampling conditions is shown in Fig. 12. A sampler tube is mounted on the
gimbaling apparatus, and is aligned with the previously measured local
vector direction of the air flow. Gas is sucked into the tube at a rate
giving a gas entry velocity corresponding to the previously measured local
gas velocity. The air is sucked by means of an aspirator. The rate of
air movement into the probe is indicated by an in-line rotameter, and con-
trolled by a small valve. Oil that enters the sampler tube along with the
air constitutes the desired sample, and it is separated out by a miniature
cyclone separator located in the sample line upstream of the air rotameter.
The small cyclone separator was very efficient at removing oil spray from
the airstream, as indicated by cleanliness of the gas exiting from the
cyclone. The oil collected by the cyclone separator flowed into a 2-cc
pipette, plugged at one end so that the oil collected therein. The oil-
sampling rate was determined by timing the rate of oil accumulation in the
pipette after a quasi-steady-state condition was established.
The effects of nonisokinetic sampling on liquid collection rate were in-
vestigated. The results for one particular location in the flowfield of
a 55-J burner are shown in Fig. 13. At the location in question, the local
free stream air velocity was 14 ft/sec, and oil sampling was performed at
velocities from 6 to 30 ft/sec. The data shown in Fig. 13 indicate that
sampling at a velocity which is too low (i.e., lower than local air
velocity) results in larger magnitude liquid collection rate errors than
sampling at velocities which are too high. For example, sampling at a
velocity of 6 ft/sec too high results in a liquid collection rate 4 per-
cent too high, while sampling at a velocity 6 ft/sec too low results in a
liquid collection rate 21 percent too low. The experiments relating to
Fig.13 utilized an 0.75-gph oil burner nozzle which produces droplets hav-
ing a mass median droplet diameter of about 75 microns. Most oil burners
produce approximately this same size or larger droplets. Since noniso-
kinetic sampling errors tend to be most significant for small droplets,
the errors represented by Fig. 13 can be interpreted as about the most
severe that normally would be encountered because of nonisokinetic effects.
36
-------
•^vVV-i. Oil .Spray
;''• ?',.'••.'..'•/. -*• Air Flow
'•'•;•'.\:~\.." • •
''
Sampler Tube
Calibrated
Class Tube
Oil Outlet
Flex Tube
Air Outlet
Flex Tube
Air Flow
Control Valve
To Aspirator
\_
Rotameter
Air Flow Indicator
Figure 12. Liquid spray mass flux sampling apparatus
37
-------
00
(0
«
12
10
8
i
Local Free Stream
Velocity
10
15
20
25
30
Sampling gas velocity, ft/sec
Figure 13. Effects of sampling velocity on oil collection rate
-------
Oil mist and vapors that arise from cold-flowing burners (without ignition,
for oil spray collection) were found to be irritating to personnel. To
minimize exposure of personnel, a plastic curtain was hung around the oil
burner under test, and air flow from behind the curtain was ducted through
a blower and directed 30 feet up the adjacent hillside. The apparatus is
shown schematically in Fig. 14. Air exhausting from the end of the duct
appeared to be free of fuel mist, and irritation of operating personnel
was eliminated. The oil sampler and the cyclone separator were mounted
on the gimbal apparatus inside the curtain.
COMBUSTION GAS FLOW PATTERNS
Burner geometry effects were characterized by air flow pattern determina-
tion in the absence of oil spray and combustion. To characterize the ef-
fects of combustion on the flow patterns, the same experiments were re-
peated, except that they were conducted with combustion. The apparatus
used for these experiments was fundamentally the same as used for the ex-
periments involving air flow only, except that several modifications were
required to make the system compatible with oil spray and combustion.
One important modification required for measurements under hot-fire condi-
tions was* the use of cooled probes. Shown schematically in Fig. 15 are
the water-cooled versions of the multiport impact probe for velocity di-
rection determinations, and a separate water-cooled pitot probe for veloc-
ity magnitude determination.
The burners were fired using the same coaxial, cylindrical sheet-metal
combustion chambers previously described. However, for the hot-fire ex-
periments, the burners were fired vertically upward rather than downward.
The upward orientation, as shown in Fig. 16, had two advantages: (1) hot
combustion gases were allowed to rise without contacting the burner hard-
ware and (2) oil drainage out of the probes was promoted. The latter is
desirable since it is important to keep oil from entering the Baratron
39
-------
Ducted 30 Ft. up
Adjacent Hillside
Plastic
Curtain
Motor
Air
Pump
Oil Burner
55-Gal.
Oil Reserve
Catch Tank
Figure 14. Oil burner cold-flow facility schematic
-------
Water
Inlet
Water
Outlet
h-V1-
18"
Water-cooled four-tube directional probe
3/8" Die
SECTION
Water
Inlet
Dynnrdc Pressure
To B/vratron
Pressure ^__
Head
Water
Outlet
Static Pressure
To Barntron
Pressure Head
Static Pressure Ports
Water-cooled pitot probe
Figure 15. Water-cooled probes
-------
Gaseous
Nitrogen
Oil Trap
(14- Avail.)
*>.
tsj
Oil Burner
Main Switch
Water-Cooled
Flame Probe
Cylindrical
Chamber
55-Gal Oil
Reservoir
Figure 16. Schematic of oil burner flame gas velocity probe system
-------
pressure-sensing head which, for electronic stability, is thermostati-
cally maintained at an elevated temperature. The composite test system
is shown in Fig. 16. Included are electrical purge/sense valves which
allow periodic purging of the probes and lines to clear any oil or soot
deposits from the probe lines. Also included are oil traps to inhibit
any possible oil migration to the Baratron pressure-sensing heads.
At the time the combustion gas flow pattern measurements were being
made, there was no provision for analysis of the gases as a function of
the location at which the impact pressures were measured. Calculation of
velocity requires knowledge of the gas density (see Eq. 1). Therefore,
in the absence of local combustion gas analyses, the combustion gas was
assumed to have a density corresponding to a molecular weight of 29 and
a temperature of 2300 F. These assumptions were used in the calculation
of all combustion gas velocities reported herein. Localized combustion
gas compositions were determined at a later stage in the program, but not
at a sufficient number of locations to include all of the velocity, measure-
ments previously taken and, hence, the composition measurements were not
incorporated into calculation of any of the combustion gas velocities re-
ported herein.
COMBUSTION CHAMBERS
There were four research cylindrical combustion chambers used in the com-
bustion gas and exhaust gas analysis studies. The four chambers were of
8-, 11-, 18-, and 30-inch diameters for a hot-firing oil flowrate range of
0.50 to 12.0 gph.
The 8-inch-ID chamber was fabricated to accommodate oil burners of 0.50-
to 1.50-gph flowrate, with oil spray-chamber wall impingement being the
primary limiting factor. Figure 17 is a drawing of the 8-inch-diameter,
cylindrical combustor showing the various locations of the sampling probes,
the semicircular mixing baffles, and a perpendicular burner port extension.
In the coaxial chamber configuration, the distance from the burner port to
43
-------
SAMPLE
COAXIAL
BURNER PORT
FLANGE
EXHAUST GAS
SAMPLING PROBE
h" PYROFLEX
INSULATING LINER
COMBUSTION GAS
SAMPLING PROBE
SEMICIRCULAR
MIXING BAFFLES
(7 PLACES)
PERPENDICULAR
BURNER PORT
EXTENSION
Figure 17. Schematic of the 8.0-inch-diameter insulated cylindrical combustor
with semicircular mixing baffles
-------
the first baffle (combustion chamber length) is 16 inches, with a total
of 7 semicircular mixing baffles spaced 6 inches apart down the remaining
length of the chamber. These baffles are installed 0.50 inch short of
either side of the chamber centerline, leaving a 1-inch gap down the cen-
ter through which the flame can be observed from the end of the combustion
chamber. The combustion chamber section is completely insulated with
1/2-inch-thick Pyroflex liner, which is a vacuum-cast alumina silicate
fiber matting that is flexible when purchased, but sets up firmly after
being fired for approximately 15 minutes. The insulating characteristics
of the Pyroflex lining maintains the inside wall temperature of the com-
bustion chamber section at about 3000 to 3200 F. This extremely high
wall temperature minimizes soot formation on the wall since the carbon
oxidizes off at this temperature. Three combustion zone sampling ports
are located on the combustion chamber section to allow local combustion
gas sampling traverses at 1.5, 3 and 6 inches downstream from the burner
head. An additional end probe is mounted near the exit of the combustor
to sample the completely combusted, mixed exhaust gases to obtain overall
conditions. The end probe is inserted approximately 12 inches in from
the exit, past the exit baffle to avoid ambient air ingestion into the
probe that would dilute the sample. The probes used for both the combus-
tion gas and exhaust gas sampling were water-cooled United Sensor Corpora-
tion Model GB 24-125. A water jacket, 19 inches in length, could also be
attached to the combustion chamber section and, with the removal of the
Pyroflex insulation, the inside wall temperature could be lowered to about
350 to 450 F to simulate a water-base (cold-wall) type furnace. The water-
jacket enclosure was sealed on the burner port end flange and open at the
other end, permitting cold-wall operation only when the chamber assembly
is mounted vertically.
An 8-inch-diameter, perpendicular burner port combustion extension, also
shown in Fig. 17, was fabricated to mate to the 8-inch-diameter cylindrical
combustor. The perpendicular burner port configuration is quite common in
commercially available furnaces, and the dimensions of this enclosure
45
-------
closely duplicated the Lennox furnace combustion chamber (described later
in this section) that was also used in experiments. The perpendicular
port extension added 11 inches to the length of the combustion section
but, being perpendicular, the closest obstruction to the flame was the
opposing wall 8 inches away. Pyroflex insulation (1/2-inch thick) was
used to line the inside of the extension, and no wall temperature varia-
tion experiments were attempted with this combustor configuration.
An 11-inch-diameter, hot-wall, cylindrical, coaxial combustor was fabri-
cated for combustion zone gas sampling experiments for the larger (6 and
12 gph) burners. The size of this combustor was marginally acceptable
for experiments at the 12-gph flowrate, as some fuel spray impingement on
the chamber wall was noted. This 11-inch combustor was similar to the
basic configuration of the smaller 8-inch-diameter combustor described
previously, with 2 probe-mounting locations (combustion section and ex-
haust), coaxial port, and straight cylindrical geometry. The mixing
baffles differed in that, instead of semicircular plates, it used tur-
bulence generators in the form of a convergent section (choke) and a
1 in. sq mesh grid of 0.25-inch water-cooled tubing. The overall length
of this combustion chamber was 7 feet, and the first 22 inches were in-
sulated with Pyroflex liner making it a hot-wall combustion chamber with
an inside wall temperature of approximately 3000 F. The combustion cham-
ber section had 2 additional probe ports (5 total) allowing combustion
gas-sampling traverses at 1.5, 3, 6, 9, and 12 inches from the oil burner
head.
A 30-inch ID, hot/cold wall, coaxial combustor was fabricated for the 6-
to 12-gph oil burner experiments. Figure 18 is a schematic of the 3-
section combustor, showing the hot- or cold-wall combustion chamber section
and the 2-piece mixing/exhaust section. The combustion chamber section
consisted of a 30-inch-diameter chamber with a 31.5-inch-diameter water-
jacket shroud surrounding it, leaving a 0.75-inch, water coolant passage
with 3 inlet and 3 outlet fittings. The water jacket was sectioned by an
46
-------
WATER JACKET SHROUD
COMBUSTION CHAMBER WALL
OBSERVATION
WINDOW
COAXIAL BURNER
PORT
WATER INLET/OUTLET
FITTING (6)
EXHAUST
SECTION
SEMICIRCULAR
MIXING BAFFLES (3)
-CHOKE RING BAFFLE (2)
.COMBUSTION ZONE
PROBE PORTS (5)
Figure 18. Schematic of 30-inch-diameter, hot/cold wall, coaxial cylindrical
research combustor for 6-to 12-gph oil burners
-------
expansion joint to compensate for differential expansion between the
jacket wall and the hotter combustion chamber wall where the wall temper-
ature differential may be as great as 1000 F in the "hot-wall" mode.
There are 5 combustion gas sampling ports that permit sampling traverses
at 3, 6, 9, 12 and 15 inches from the burner head. The ports were
bellowed tubes welded at the inner and outer cylinder walls to prevent
water leakage and allow differential expansion of the walls. The combus-
tion chamber can be run in the "cold-wall" mode by flowing water coolant
(5 to 6 gpm at 12 gph firing rate) through the water jacket, which re-
sults in an inside chamber wall temperature of about 350 to 400 F. The
"hot-wall" configuration required the installation of 1/2-inch Pyroflex
lining held in place by a water-cooled support grid of 1/4-inch stainless-
steel tubing. A trickle flow of water must be maintained through the
waterjacket to prevent the inner cylinder wall from overheating (T =
1500 F). The combustion section was 36 inches in length, terminating with
a choke ring-type mixing baffle of 22 inches ID. Downstream of the choke
ring were the mixing/exhaust sections consisting of 2 semicircular mixing
baffles and another choke ring baffle, spaced at 12-inch intervals. The
exhaust was turned 90 degrees vertically and terminated near the exit with
another semicircular baffle to avoid ambient air ingestion that would
dilute the gas samples. The sample probe extends 18 inches into the ex-
haust section to further prevent sample dilution. A 5-inch-diameter Pyrex
glass window in the burner port flange allowed observation of the flame in
the combustion chamber.
An 18-inch-diameter, hot-wall, coaxial combustor extension to the 30-inch-
diameter mixing/exhaust section was fabricated to allow 3-gph burner ex-
periments with a minimal of test stand modifications. The combustion
chamber was 18 inches in length, lined with 1/2-inch-thick Pyroflex in-
sulation, and terminated with a 12.5-inch-ID choke ring baffle/mating
flange combination. The chamber was designed for hot-wall experiments
only, with no provisions for water cooling the wall. There were 4 com-
bustion zone probe ports that allowed sampling traverses to be made at
1.5, 3, 6, and 9 inches from the burner head. The chamber mated directly
48
-------
to the 30-inch-diameter mixing section and required only a partial cover-
ing of the exhaust outlet to avoid air ingestion at the lower firing rate.
Figure 19 shows a schematic of the general test stand setup of both the
8-inch- and 30-inch-diameter combustors. This is a schematic showing
general locations rather than detail of flow systems, and does not show
the combustion zone probe locations.
FURNACES
Two commercially available, home-sized (0.5 to 1.50 gph), heating furnaces
were used in oil burner flue gas sampling experiments. A Lennox Model
OF7-105M, 0.75-gph warm-air furnace and a Unitron Model A100, 1.10-gph
combination warm-air/hydronic (warm water) furnace were used to test the
0.5- to 1.50-gph oil burners.
The Lennox warm-air furnace had an 8-inch-ID by 14-inch-tall upright circu-
lar combustion chamber with the burner port mounted perpendicular to the
chamber axis. The chamber was refractory-lined (hot wall) with the .heat
exchanger section mounted directly above the combustion chamber. The
combustion chamber configuration was very similar to the 8-inch-diameter,
perpendicular port, research combustor extension designed from basic di-
mensions taken from the Lennox furnace. The flue samples were removed
from the 6-inch-diameter stack through an uncooled, 1/4-inch-OD, stain-
less-steel tube having its entrance bent perpendicular into the direction
of the gas flow.
The Unitron Model A100 warm-air/hydronic furnace is rather unique in its
combination heat transfer system, using both forced air and hot water for
heat distribution. This furnace had a coaxial combustion chamber very
similar to the 8-inch-diameter research combustor described earlier in
this section. The furnace combustion chamber is cylindrical, 10.5-inch
ID, 24 inches in length, and lined with 1/2-inch Pyroflex insulation. The
49
-------
*8-IN. DIAMETER
COMBUSTOR
EXHAUST GAS
SAMPLING PROBE"7
tn
O
NO. 2
FUEL OIL
RESERVOIR
FUEL
SYSTEM SHUTOFF
VALVE
6-12 GPH
OIL BURNER
0.5-1.5 GPH-
OIL BURNER
30-IN.
DIAMETER
OMBUSTOR
BURNER ON-OFF
SWITCH
TO GAS ANALYSIS
INSTRUMENTS
Figure 19. Schematic of oil burner combustion gas sampling system
-------
hot combustion gases are then turned 90 degrees upward into a manifold
and 90 degrees again through a multitube hot gas-to-water heat exchanger
to heat the water, which is the primary heat transfer medium. Part of
the hot water flows through an internal, finned radiator which, with the
blower, constitutes the warm-air heat distribution system.
COMBUSTION GAS COMPOSITION PATTERNS
As a tool for diagnosis to determine the local combustion effects leading
to formation of the various pollutants, it is desirable to know not only
the local flow patterns, but also the local composition patterns of the
combustion gases. To accomplish this required localized sampling and
analysis of combustion gases. The localized sampling was conducted using
the combustion chambers having probe ports, as described in the previous
section.
Accurate measurement of nitric oxide concentration by gas sampling from a
gas stream sufficiently hot to allow non-negligible nitric oxide chemical
kinetics requires that the sample probe instantaneously quench the temper-
ature of the gas sample as it enters the probe. If the temperature is
not instantly quenched to a temperature sufficiently low so that nitric
oxide kinetics are negligible, there is a definite probability that the
nitric oxide content of the gas sample will change within the probe as a
result of chemical processes on surfaces and in the bulk gas flowing
through the probe. For the present program of study, the likelihood of
quantitative changes in nitric oxide concentration due to sample quench-
ing technique was recognized; however, the nature of this study (pollutant
minimization) was such that exact concentrations of nitric oxide are of
considerably less importance than relative concentrations at various lo-
cations and for various burners. It was chosen, therefore, to utilize a
simple water-cooled convective quench sample probe rather than a more
costly, difficult-to-operate probe designed specifically to achieve rapid
thermal quench.
51
-------
The gas sample probe utilized to obtain gas samples from within the
burner combustion zone was a low-cost, commercially available model
(GB 24-125) manufactured by United Sensor Corporation. A photograph of
this probe is shown in Fig. 20. It is a simple tubular probe with a
short right-angle turn at the sampling end. The right-angle end allows
the probe to be inserted from a position at right angles to the flow,
with the probe inlet still being pointed upstream. No attempt was made
to align the gas sample probe inlet with the local vector direction of
the gas flow. The gas velocity within the 1/8-inch-diameter inlet to the
probe was maintained relatively constant for all experiments at approxi-
mately 250 ft/sec.
In one instance, the velocity was reduced by a factor of 4:1, changing
the thermal quench conditions, and the nitric oxide analysis was found to
be insensitive to this change.
The sample flow train used for the experiments is shown in Fig. 21. Gas
aspirated through the sample probe entered an air-cooled condensibles
trap where particulates and heavy oils were separated out. Next, the gas
passed into an ice-cooled, stainless-steel condensibles trap where most of
the water and all but the most volatile hydrocarbons were removed. After
the condenser, the gas passed into a pyrex wool-filled glass cylinder.
The last item served as a final separator for heavy oils and particulates,
and provided a visual indication of the cleanliness of the gas being ad-
mitted to the analysis instruments. Table 2 gives a summary of the gas
analysis instruments used. The gas leaving the glass wool filter was
split into three parallel paths. One path led directly to the total hy-
drocarbon analyzer. A second path led through a Drierite bed where water
was removed, and then into the series-plumbed carbon monoxide, carbon
dioxide, and oxygen analyzers. The third path passed through a combined
Drierite and 3 A molecular sieve bed for total water removal, and then
into the nitric oxide analyzer. The gas was pumped through the system by
three diaphragm pumps located downstream of the nitric oxide analyzer,
total hydrocarbon, and carbon monoxide + carbon dioxide analyzers. The
52
-------
.,!., . ' -~ - ;;
Figure 20. United Sensor model GB 24-125 gas sampling probe
53
-------
INSULATED
LINE
GLASS
WOOL
FILTER
GLASS
CONDENSIBLES
TRAP AIR-COOLED
SIGHT
TUBE
GLASS
STAINLESS STEEL
CONDENSIBLES TRIP
ICE-COOLED
MOLECULAR SIEVE 3A*
+ INDICATING DRIERITE BED
3-WAY SELECTOR
DILUENT ^ r-. ,-,_
.....
AIR PURIFIERS
— ^
NO
X
ANALYZER
TOTAL
HC
ANALYZER
DUAL DILUTION
SYSTEM
DRIERITE
BED
SAMPLE STREAM FLOW ».
ALL LINES 1A-INCH STAINLESS STEEL TUBING (THIN WALL)
CO
ANALYZER
€
co
ANALYZER
VENT
TO
HOOD
\NALYZER
Figure 21. Analytical system for fuel oil burner emissions analysis
-------
Table 2. EXHAUST ANALYSIS INSTRUMENTS
Type
Range
Sensitivity
Calibration
CO
MSA
Nondispersive IR
LIRA
Model 300
0 to 1500 ppm
fnole)
30 ppm minimum
detectable
1000 ppm CO in
N_ standard gas
co2
MSA
Nondispersive IR
LIRA
Model 300
0 to 20 mole %
0.25% minimum
detectable
14% C02 in N2
standard gas
NO
MSA
Nondispersive IR
LIRA
Model 200
0 to 500 ppm
(mole)
10 ppm minimum
detectable
0.82% C.H in
N, used as
simulant for
410-ppm NO
standard
Total HC
MSA
H2 flame
ionization
detector
0.2 to 800 ppm
total HC by
volume as CH^
10 ppm minimum
detectable
3% CH4 in helium
used as a
standard
Oxygen
Beckman
polarographic
0 to 100%
-0.1%
Air - 21%
N2 = 0%
Smoke
Bacharach
(manual)
0 to 9
1
Ten spots of
monotonically
varying
darkness
en
en
-------
system also contained capability for dilution of gas passing through the
carbon monoxide, carbon dioxide, and oxygen path. Dilution was achieved
by admission of air metered through parallel rotameters, and it allowed
use of the analytical instruments on gas samples more concentrated than
the highest range of the instruments.
When the analytical system shown in Fig. 21 is used to analyze gases sam-
pled from within the combustion zone, there are two major factors that
must be considered when reducing the data: (1) only burned or partly
pyrolyzed fuel is included in the analysis, since liquid or vaporized fuel
is non-quantitatively removed by the cold trap, and (2) water formed from
hydrogen and oxygen during the combustion process is also removed by the
cold trap from the analyzed sample.
Values calculated from the local combustion gas sampling data include:
the local stoichiometric ratio of the gases (this necessarily excludes
unburned fuel vapor or liquid which is removed at the cold trap), the
weight content of nitric oxide per unit weight content of burned fuel,
and the weight of carbon monoxide per unit weight of burned fuel. The
method of calculation to obtain these values is described below.
The calculations are based on air having the nominal composition given in
Table 3.
Table 3. PURE AIR COMPOSITION
Component Mole Percent
N2
°2
Ar + He + Ne
C02
100.00
56
-------
The composition of the fuel is assumed to be characterized by the formula
CHX where, for use herein, x = 1.814. The following symbols are used in
the calculations:
AIR = moles of air to produce 100 moles of dry sample
FUEL = moles of fuel to produce 100 moles of dry sample
CO = moles of carbon monoxide in 100 miles of dry sample
C02 = moles carbon dioxide in 100 moles of dry sample
NO = moles of nitric oxide in 100 moles of dry sample
02 = moles of oxygen in 100 moles of dry sample
HC = moles of hydrocarbon, as Qfy, in 100 moles of dry
sample
x = number 2 fuel oil composition, CHX, xc: 1.814
The values of CO, O>2, NO, 02, and HC are obtained directly from the anal-
ysis instruments. In the following, it is assumed that all hydrogen is
oxidized to water and condensed out of the system at the cold trap, prior
to analysis.
An oxygen balance yields:
0.2095 AIR = C02 - 0.0003 AIR + O.S CO + 0.25X (C02 + CO -
0.0003 AIR] + 0.5 NO + 02
(2)
The left-hand side of the above equation represents the total free oxygen
contributed by the air. The first two items on the right side represent
moles of oxygen tied up in C02, less the amount of C02 originally present
in the air. The third term represents moles of oxygen tied up as carbon
monoxide. The fourth term represents oxygen consumed to oxidize hydrogen
yielding the water condensed out in the cold trap. The fifth term is the
oxygen tied up in nitric oxide. The sixth term is free oxygen remaining
in the sample reaching the analysis instruments. Equation 2 can be ar-
ranged to yield:
(1 + f) CO, + (1/2 + J) CO + 1/2 NO + 07
2 r _L ~. r \
0.2095 + 0.0003 +" 0.0003 x7'4^3J
57
-------
A carbon balance can be used to calculate the moles of burned fuel per
100 moles of dry sample gas:
FUEL = C02 - 0.0003 AIR + CO (4)
The moles of air available per mole of burned fuel in the sample gas can
be obtained by taking the ratio of the values from Eq. 3 and 4. AIR
must be calculated first, before calculation of FUEL. If the combustion
were in stoichiometric proportions, the moles of air would be, by an oxy-
gen demand calculation:
_ (1 * x/4) FUEL (
stoich 0.2095 li>J
The stoichiometric ratio of the locally sampled burned gases is a param-
eter frequently used in this report. It is defined as the ratio of AIR
to AIRstoich:
SR = stoichiometric ratio = 775 (6)
AIRstoich
Combination of Eq. 3 through 6 yields a direct calculation of the burned
gas stoichiometric ratio in terms of the measured parameters:
(1 + j) C02 + (1/2 + J) CO + 1/2 NO + 02
_ 0.2095 t- 0.0003 * 0.0003x/4
(1 + J) f" (1 + j) C02 + (1/2 x J) CO + 1/2 NO + 02~|
0.209T C°2 + C0 ' °-0003 0.2095 + 0.0003 + 0.0003x/4I
L (7)
According to the above definition, when the sample contains just a suffi-
cient amount of air to oxidize all of the fuel in the sample to C02 plus
condensed-out water, then SR = 1; and, as a second example, if there is
twice the required amount of air for complete oxidation of the fuel, then
SR = 2. Note that the stoichiometric ratio, as calculated from Eq.7 does
not require that the product in the sample be in chemical equilibrium.
58
-------
Note that the accuracy of the stoichiometric ratio calculation would be
affected very little if all terms in Eq. 7 containing the factors 0.0003
and NO were ignored. These factors represent the carbon dioxide origin-
ally present in free air, and the oxygen tied up in nitric oxide,
respectively.
One partially questionable assumption made in the formulation of Eq. 7
was that all hydrogen originally present in the fuel becomes oxidized to
water and removed in the cold trap. This was a necessary assumption,
since there was no instrument available to measure the actual hydrogen
content of the sample gas. The assumption is very good under the combined
conditions of air-rich stoichiometric ratios (SR > 1) and chemical equi-
librium. To test this assumption, the Rocketdyne thermochemical computer
code was used to calculate the species concentrations under conditions of
chemical equilibrium for stoichiometric ratios from 0.8 to 2.8. These
calculations included the equilibrium presence of free fy- The actual
stoichiometric ratios of these combustion gases, compared to the stoichio-
metric ratios calculated from Eq. 7 (which does not recognize the presence
of H2) are given in Table 4, where it can be seen that Eq. 7 is quite ac-
curate except for SR < 1. The calculated equilibrium conditions for these
results are given in Tables 5 and 6.
TABLE 4. VALIDITY OF STOICHIOMETRIC RATIO CALCULATIONS OF EQ. 7
Stoichiometric Ratio Calculated
Actual Stoichiometric Ratio from Eq. 7
0.800 0.844
1.000 1.003
1.200 1.197
1.400 1.400
1.600 1.600
2.000 2.002
2.400 2.404
2.800 2.804
59
-------
Table 5. EQUILIBRIUM FLAME PROPERTIES FOR NO. 2 DISTILLATE FUEL OIL
(CH. aiA, 18,443 Btu/lb Heat of Combustion With Air at 14.67 psia)
J. • O l^T
Stoich.
Ratio*
0.8
1.0 1
1.2 j
I
1.4 '
1.6 Air
2.0 Rich
2.4 1
2.8 T
0.8
1.0
1.2
1.4
1.6
2.0
2.4
2.8
0.8
1.0
1.2
1.4
1.6
2.0
2.4
2.8
Oil + Air
Inlet Temp. ,
F
0
70
200
Flame
Temperature ,
F
3429
3614
3290
2940
2649
2209
1897
1663
3778
3649
3336
2991
2703
2765
1955
1722
3567
3709
3418
3085
2802
2369
2061
1831
cp
Frozen,
Btu/lb-R
0.346
0.341
0.333
0.324
0.318
0.307
0.298
0.291
0.347
0.341
0.333
0.325
0.318
0.308
0.299
0.193
0.347
0.342
0.334
0.326
0.320
0.309
0.301
0.295
Y
Frozen
1.261
1.254
1.260
1.267
1.275
1.288
1.298
1.308
1.261
1.254
1.259
1.267
1.274
1.286
1.297
1.306
1.260
1.257
1.259
1.266
1.273
1.284
1.294
1.305
Viscosity,
centipoise
0.0666
0.0687
0.0653
0.0615
0.0581
0.0527
0.0487
0.0456
0.0671
0.0691
0.0658
0.0621
0.0589
0.0535
0.0495
0.0464
0.0681
0.0698
0.0668
0.0632
0.0600
0.0548
0.0509
0.0479
Thermal
Conductivity,
Btu/hr-ft-F
0.0702
0.0711
0.0661
0.0610
0.0567
0.0500
0.0452
0.0415
0.0709
0.0715
0.0667
0.0617
0.0574
0.0509
0.0461
0.0425
0.0720
0.0725
0.0678
0.0629
0.0588
0.0524
0.0477
0.0441
Prandtl
Number
0.7946
0.7984
0.7954
0.7915
0.7880
0.7820
0.7771
0.7730
0.7948
0.7984
0.7956
0.7918
0.7884
0.7825
0.7778
0.7738
0.7951
0.7983
0.7958
0.7925
0.7890
0.7834
0.7790
O."751
Molecular
Weight
27.73
28.80
29.00
29.03
29.03
29.02
29.01
29.00
27.72
28.77
29.00
29.03
29.03
29.02
29.01
29.00
27.71
28.73
28.98
29.02
29.02
29.02
29.01
29.00
*Stoichiometnc ratio is unity at 14.49 pounds of air per pound of fuel, and proportionate!)
greater than unity for increasing amounts of air; S.R. = (AIR/FtJF.L)/14.49
-------
Table 6. CALCULATED EQUILIBRIUM GAS COMPOSITION, MOLE PERCENT
Stoich.
Ratio
0.8
1.0
1.2
1.4
1.6
2.0
2.4
2.8
0.8
1.0
1.2
1.4
1.6
2.0
2.4
2.8
0.8
1.0
1.2
1.4
1.6
2.0
2.4
2.8
Oil + Air
Inlet Teirp. , F
0
70
200
H
0.0630
0.0397
0.000
0.000
0.000
0.000
0.000
0.000
0.0737
0.0455
0.0000
0.0000
0.0000
0 . 0000
0.0000
0.0000
0.0964
0.0577
0.0000
0.0000
0.0000
0.0000
0 . 0000
0.0000
0
0.0000
0.0313
0.0217
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0362
0.0261
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0468
0.0356
0.0000
0.0000
0.0000
0.0000
0.0000
Ar
0.821
0.866
0.882
0.890
0.895
0.902
0.907
0.910
0.821
0.866
0.882
0.890
0.895
0.902
0.907
0.910
0.821
0.864
0.882
0.890
0.895
0.902
0.907
0.910
OH
0.0499
0.2816
0.1862
0.0757
0.0790
0.000
0.000
0.000
0.0613
0.3072
0.2082
0.0885
0.0351
0.000
0.000
0.000
0.0878
0.3579
0.2533
0.1157
0.0493
0.0000
0.0000
0.0000
H2
2.016
0.250
0.030
0.000
0.000
0.000
0.000
0.000
1.996
0.269
0.036
0.000
0.000
0.000
0.000
0.000
1.964
0.304
0.048
0.000
n.ooo
0.000
0.000
0.000
H20
12.263
11.690
10.141
8.832
7.799
6.297
5.276
4.541
12.271
11.647
10.121
8.824
7.795
6.297
5.276
4.541
12.273
11.562
10.078
8.806
7.787
6 . 295
5 . 276
4.541
CO
7.243
1.393
0.161
0.0203
0.000
0.000
0.000
0.000
7.268
1.501
0.195
0.026
0.000
0.000
0.000
0.000
7.318
1.710
0.270
0.042
0.000
0.000
0.000
0.000
co2
8.687
12.052
11.247
9.841
8.679
7.000
5.864
5.046
8.659
11.934
11.210
9.835
8.678
7.000
5.863
5.046
S.604
11.705
11.127
9.816
8.672
7.0PO
5.863
5.046
NO
0.000
0.253
0.390
0.2955
0.2080
0.0829
0.0339
0.000
0.017
0.272
0.404
0.322
0.223
0.096
0.041
0.018
0.027
0.310
0.451
0.373
0.268
0.125
0.059
0.028
N2
68.837
72.522
73.784
74.465
74.947
75.603
76.028
76.326
68.901
72.456
73.751
74.447
74.933
75.596
76.023
76.323
68.796
73.328
75.683
74.405
74.905
75.582
76.015
76.319
°2
0.000
0.619
3.160
5.566
7.444
10.107
11.888
13.161
0.000
0.666
3.159
5.553
7.432
10.100
11.SS4
13.159
0.000
0.754
5.162
5.526
7.406
10.085
11.876
13.154
-------
The primary cause of the inaccuracy at SR < 1 is the presence of H2- In
nonequilibrium gases, there is likely to be H2 present even where none
would be indicated from equilibrium calculations and, at fuel-rich condi-
tions, there could be more or less than indicated from the equilibrium
calculations. Because of this likelihood of nonequilibrium, no attempt
was made to correct the calculations of Eq. 7 by means of equilibrium
calculations.
The other parameters of interest for the sampled combustion gases are the
mass ratio of nitric oxide to burned fuel, the mass ratio of carbon mon-
oxide to burned fuel, and the mass ratio of unburned hydrocarbons (as CH4)
to burned fuel. These ratios are generally expressed herein as grams of
nitric oxide per kilogram of burned fuel (gm NO/kg fuel) grams of methane
per kilogram of fuel (gm Gty/kg fuel), and grams of carbon monoxide per
kilogram of burned fuel (gm CO/kg fuel). These parameters are calculated
by aid of Eq. 3 and 4 from the following relations:
un 100° x NO x MWMn
gm NO NO ,-g-.
kg fuel = "(C02 - 0.0003 AIR + CO) MWp *• J
CO 1000 x CO x MWCQ
kg fuel (C02 - 0.0003 AIR +
„_ 1000 x HC x MWru
em HC L"
kg fuel (C02 - 0.0003 AIR + CO) MWF
where
MWNQ = molecular weight of NO = 30.01
MWp = molecular weight of fuel
= 12.01 + 1.01 x = 13.83
MWCO = raolecular weight of CO = 28.01
MWru = molecular weight of methane = 16.04
L»4
For calculation of the above quantities, the term 0.0003 air can be as-
sumed zero without introducing more than about 0.1-percent error in the
,101
l±UJ
62
-------
calculations, or AIR can be computed from Eq. 3 and included in the cal-
culation. The numbers given in this report include the effect of the
term. The experimental data were reduced, according to the above equa-
tion, by means of a remote terminal timeshare computer program.
EXHAUST GAS ANALYSES
In several instances, the totally mixed exhaust gases from oil burner
combustion processes were analyzed. These measurements can be used to
compute the total emissions of the burner. The mixed exhaust gases were
analyzed: (1) in the exhaust stack from the furnaces used, and (2) down-
stream of the mixing baffles of the research combustion chambers described
previously.
When the exhaust gases were sampled from the exhaust stack of the furnaces,
the gases were, of course, already cool (approximately 600 F) from passage
through the furnace heat exchanger. Since the flue gases were precooled,
the sample for analysis was withdrawn through an uncooled, 1/4-inch-OD,
stainless-steel tube with its entrance perpendicular to the direction of
gas flow. The gas withdrawn through this uncooled tube was then input to
the sample analysis system shown in Fig. 21.
When the exhaust gases were sampled from the end of the mixing section of
one of the previously described research combustors, the gases were still
relatively hot due to the absence of a heat exchanger. Therefore, the
sample gases were removed by means of the convective quench probe shown
in Fig. 20. From the probe, the gases passed to the sample analysis sys-
tem shown in Fig. 21.
The parameters calculated for the exhaust gases were stoichiometric ratio,
nitric oxide-to-burned fuel mass ratio, carbon monoxide-to-burned fuel
mass ratio, and unburned hydrocarbon-to-burned fuel mass ratio. These
parameters were calculated from Eq. 7 through 10.
63
-------
In addition to the gaseous pollutants described above, the smoke content
of the mixed gases was also measured. The instrument utilized for this
purpose was a Bacharach smoke meter. It is manufactured by the Bacharach
Instrument Company, Pittsburgh, Pennsylvania. This is a hand-held device
which, when pumped, sucks flue gases from a 1/4-inch-OD, uncooled sample
probe through a piece of white filter paper; 10 strokes of the pump, over
a period of about 15 seconds, causes the passage of 2250 in.3 of flue gas
per in.2 of filter paper. The smoke particles deposit out on the filter
paper. A reading is taken by comparing the darkness of the smoke deposi-
tion spot to a scale of 10 such calibrated spots provided with the instru-
ment. The readings vary from 0 to 9. A reading of zero corresponds to
no visually detectable deposition on the filter paper, while a reading of
9 corresponds to a dark black deposition. Intermediate readings are vary-
ing shades of black and gray, increasing in darkness with increasing read-
ing numbers. A reading of 1 is generally accepted by the industry as a
very acceptable degree of smoke. At the opposite extreme, a reading of
9, which is totally unacceptable, still does not correspond to sufficient
smoke to be easily visible from observation of the exhaust stack outlet.
In some instances, reported herein, when the reading was obviously greater
than 9, the number of strokes was halved and smoke spot reading doubled, thus
extending the smoke scale to a maximum of 18.
THERMAL RADIATION
Several experiments were performed with an instrument for measuring thermal
radiation. This instrument is shown in Fig. 22. The detection end of the
radiometer probe has a small hole, which is at one focus of an ellipsoidal
polished cavity. At the other focus of the cavity is a hemispherical de-
tector which acts as a blackbody to soak up the thermal radiation. Heat
is conducted away from the blackbody detector through a small metal rod.
Thermocouples at two locations on the metal rod detect thermal gradients
therein, which are an indication of the heat flux through the rod and,
hence, an indication of radiant heat flux impinging on the blackbody de-
tector. The ellipsoidal cavity was purged continuously with gaseous
64
-------
Figure 22. Land 2i\ radiometer used for radiation measurements
65
-------
nitrogen during use of the radiometer to keep possibly harmful combus-
tion gases from entering the polished ellipsoidal cavity. The probe was
calibrated by the manufacturer in terms of radiation, using the specific
nitrogen purge rate recommended for use. In practice, the nitrogen purge
was apparently adequate to keep combustion gases from entering the probe
cavity, but it was not able to stop oil spray from entering the cavity.
After a few experiments, the oil spray which entered the cavity deterior-
ated the blackbody coating (apparently a paint) on the detector and null-
ified the calibration of the radiometer. The data reported with this
instrument in a later section of this report should, therefore, be con-
sidered only as semiquantitative indications.
NONCONVENTIONAL BURNERS
There were four nonconventional burners designed and fabricated to in-
vestigate either the effects on exhaust pollutants or the feasibility of
various designs utilizing known pollutant-reducing burner concepts. These
designs are called: (1) mechanically rotated, intense swirl burner, (2)
displaced oil injection burner, (3) forced recirculating combustion gas
burner, and (4) heated air burner.
Intense Swirl Burner
An existing burner was modified to accept a variable rate, mechanically
rotated, multiple-vane swirler assembly to be used to investigate low to
very high air swirl rates. Figure 23 shows a schematic of the modified
burner with the motorized swirler assembly. A six-vane swirler was
mounted on a hollow rotating shaft, through which an inner stationary tube
held the oil spray nozzle in position and also served as the oil feedline
to the nozzle. The outer shaft was pulley driven by a V-belt link to a
1/7-hp (=1/12 hp required), 1725-rpm electric motor. The 4-cluster pulley
arrangement allowed a swirl spin rate variation from 812 to 3450 rpm, vary-
ing the ratio of angular momentum to axial momentum (swirl number) of the
66
-------
4 CLUSTER
COMBINATION
V-BELT PULLEY
STATIONARY
OIL SPRAY
NOZZLE
ROTATING
6 VANE
SWIRLER
VERSATILE BURNER
BODY
CHOKE
RING
Figure 23. Schematic of modified versatile burner with variable rate,
mechanically driven, multiple vane swirler
67
-------
air from 2.2 to 8.8. The swirler assembly was mounted in a standard 4.0-
inch-diameter blast tube and utilized a 30-degree convergence, 1.65-inch
choke diameter, fixed geometry burner head (Fig. 24). Figure 25 shows
photographs of this mechanically rotated, intense swirl burner in various
stages of assembly.
Displaced Oil Injection Burner
A displaced injection experimental burner (Fig. 26) was constructed and
tested. The intent of this injection concept was to utilize cool, re-
circulated combustion gases to aid fuel vaporization and, thereby, alter
stoichiometric ratio at which combustion occurred. The primary air was
introduced tangentially into the chamber, causing a recirculation pattern,
and the oil nozzle was displaced circumferentially downstream into the
combustion chamber. This extension allowed time for the primary air and
the recirculated combustion gases to mix prior to encountering the oil
spray. The first burner (prototype A, Fig. 26a) was designed to minimize
modifications to the conventional burner. A burner was fitted with a
4-inch-diameter blast tube cut at 45 degrees at the exit with an oil noz-
zle canted 60 degrees to the blast tube centerline. The choke plate con-
figuration was established by trial through visual observation of the
flame. It was found that substantial air flow was needed near the chamber
wall to keep the fuel spray from impinging on the chamber wall and causing
smoke formation. The minimum height of the oil nozzle was dictated by the
length of the nozzle and the chosen 60-degree cant angle. This combina-
tion of requirements for the nozzle height and chamber wall air flow re-
sulted in a choke plate configuration that produced poor mixing of the
primary air and fuel. Very smoky, high NO emission profiles were obtained.
The lowest smoke reading obtained in the hot-firings with a No. 2
Bacharach smoke scale.
Modifications were made (prototype B, Fig. 26b) to decrease the oil nozzle
height (i.e., extension into the chamber) and improve the oil spray contact
68
-------
0.75" R
Figure 24. Cross-section view of the fixed geometry, 1.65-inch-
diameter, 30-degree convergent choke
69
-------
5AD26-7/17/73-S1D
5AD26-7/17/73
Figure 25. Photographs of the mechanically rotated, six-vane
intense swirl burner in various stages of assembly
70
-------
(a) Prototype A
Small Versatile
Burner Body
Perpendicular
View of Choke Plate
Space Available
i- for Primary Air
8-Inch Diameter
Combustor
Space for Bypass
Air, Determined by
Mounted Nozzle Height
(b) Prototype B
I
Space Available
for Primary Air/
Fuel Mixing
Space for Bypass Air
Figure 26. Schematic of displaced oil injection burner prototypes
constructed and hot fired
71
-------
with the primary air flow. To accomplish this, the oil nozzle was in-
serted through the chamber wall and a smaller, 2.0-inch-diameter blast
tube was used to concentrate the air flow near the wall and oil spray.
No choke plate evaluation hot firings were made and the burner was fired
with the straight 2-inch-diameter blast tube. The oil spray, now much
closer to the wall, was blown onto the wall insulation and "cooked out,"
producing excessive white, unburned fuel smoke. No further modifications
were made to improve the mixing in the displaced injection prototypes.
Recirculating Combustion Gas Burner
An experimental burner was designed and fabricated that would force mixing
between inlet air and combustion gases withdrawn from the combustion cham-
ber. This mixture was then to be used as the air supply for the burner.
The forced recirculation burner was designed so that the recirculated
combustion gases were withdrawn from portions of the chamber where heat
may have been lost from the gas to the chamber wall and, therefore, the
burner to some extent simulated flue gas (i.e., cooled gas) recirculation
without the required external ducting. The flow directions are shown
schematically in Fig. 27. Air entered at the eight side ports, flowed to
the front of the burner, where it made a 180-degree turn, simultaneously
mixing with combustion gas drawn in through the 12 end ports. The mixed
gases flowed back into the internal cavity of the squirrel cage blower
wheel (later replaced by a centrifugal blower wheel), where centrifugal
forces generated by the blower forced the gases radially outward into the
plenum from which they then flowed through holes in the back plate of the
blower wheel, into the blast tube funnel inlet and down the blast tube
toward the blast tube exit (burner head). The air flow was regulated by
a cover band that slid over the air inlet holes. The combustion gas flow
was regulated by the size of the combustion gas inlet holes, which could
be plugged or drilled out as necessary. The ignition source was external
or manually operated. Figure 28 is a photograph of the assembled burner
with a mixed-gas temperature probe (iron/constantan thermocouple) and a
mixed-gas sampling port.
72
-------
COMBUSTION
GAS ENTRY
PORTS (12)
OIL
SUPPLY
(100 PSIG)
1
i
INCHES
PRIMARY GAS
SAMPLE
OUTLET
AIR ENTRY
LUMlil.sLU
GASES
MIXED GASES
(PRIMARY GAS)
COMBINLD
CASKS
MODIFIED SQUIRREL CAGE FAN
OIL .NOZZLE AND
SKIRL VANE ASSEMBLY
PRIMARY GAS
TEMPERATURE
(IRON-CON)
Figure 27. Schematic of the forced recirculated combustion gas experimental burner
-------
(
PRIMARY GAS
THERMOCOUPLE
PRIMARY GAS
SAMPLE OUTLET
COMBUSTION GAS
INLET PORTS
MOTOR HOUSING
8-VANE, 60° SWIRLER
5AD24-6/27/73-S1A
Figure 28. Photograph of the forced recirculating combustion gas burner (fully instru-
mented) with an 8-vane, 60-degree swirler ring, and 4 of 12 combustion gas
inlet ports plugged
-------
Heated Air Burner
The heated air, oil burner concept was investigated as a part of a three-
stage combustion burner concept. The preheating of inlet air was to be
the initial conditioning for the first, fuel-rich combustion stage. The
heated air burner was designed and fabricated to investigate the effect
of the inlet air temperature on exhaust emissions, especially that of
smoke emission. The added stage of preheating the inlet air was intended
to accelerate the primary combustion process (fuel-rich combustion) by
prevaporization and thus avoid quenching the air/hydrocarbon kinetics in
the free-carbon stage, thereby inhibiting the formation of smoke. High
smoke emission levels are a common problem in two-stage combustion proc-
esses, as the free carbon from the primary stage is not easily oxidized
in the secondary combustion stage. Figure 29 shows a schematic of the
heated air burner assembly. The system utilized the Carlin 250 FFD
burner body to provide the air and oil flows, a 16-inch-diameter by 22-
inch- long, sealed air shroud, and a 2-inch-diameter burner head with an
eight-vane, 75-degree swirler ring. An iron-constantan thermocouple
mounted near the entrance to the burner head measured the preheated air
temperature which reached as high as 590 F.
OIL NOZZLES
All oil burners tested were of the high-pressure atomizing type. These
burners all used high pressure (100 to 300 psig), swirl atomizing, simplex
nozzles requiring only oil supplied at high pressure for spray atomization.
Figure 30 shows a cross-section schematic of a typical nozzle. The high-
pressure oil flows through the filter, then the flow direction is aligned
axially by the flow distributor body for a controlled entrance into the
swirl plug. The swirl plug contains a number of slots (oil flow channels)
that direct flow tangentially into the swirl cavity, imparting rotational
energy to the oil flow before it is discharged through the discharge ori-
fice. The combination of the small orifice size and high swirl
75
-------
4" Diameter
Flex Duct
8" Diameter
Uninsulated Combustor
16" Diameter
Air Shroud
Drive Burner
(Carl in 250 FFD Body)
Inlet Air
Thermocouple
Figure 29. Schematic of the preheated air burner test apparatus using a 2-inch-diameter
blast tube with an 8-vane, 75-degree swirl ring, and a 0.75-80°-C oil nozzle
-------
DISCHARGE
ORIFICE
SWIRL
CAVITY
SLOTTED
SWIRL
PLUG
FLOW
DISTRIBUTOR
BODY
SINTERED METAL
OR WIRE MESH
FILTER
TYPE A
HOLLOW CONE
TYPE B (OR R)
SOLID CONE
TYPE C (OR W)
SEMI-SOLID CONE
'SPRAY CONE ANGLE
Figure 30. Cross-sectional schematic showing the internal
construction of a typical high-pressure oil atomi-
zing nozzle and the various spray cone patterns
77
-------
(rotational energy) causes the oil to self-atomize without requiring the
addition of more energy from other sources. The resulting droplets are
very small, with the mass median dropsize (D),less than 100 microns
(1 micron = 10 meters) up to nozzles as large as 25-gph flowrate
at 100-psig oil supply pressure. Various combinations of discharge ori-
fices and swirl plugs produce a variety of spray patterns which are gen-
erally categorized into three groups: hollow cone, solid cone, and semi-
solid cone. Figure 30 shows these three spray patterns and their type of
nomenclature. The hollow cone pattern is generally recommended for the
0.5- to 3.0-gph burners and the solid cone pattern for the larger burners.
The spray cone angle is usually recommended by the burner manufacturer
for each specific burner and it is an important parameter in mixing and
also flame front stability, especially in "conventional" (nonflame-reten-
tion head) oil burners. The nozzles are labeled according to their nominal
oil flowrate (gph at 100 psig), the spray cone angle (degrees) and the
spray pattern type.
Except where otherwise noted, oil nozzles used in the experimental studies
described herein were manufactured for oil burner applications by Delavan.
78
-------
EXPERIMENTAL RESULTS
Presented in this section are discussions and results of experimental
studies conducted with the oil burner apparatus. It is separated into:
(1) Atomization Measurements, (2) Commercial Burner Studies, (3) Inter-
pretation of Commercial Burner Results, (4) Burner Geometry Op|timiza-
tion Studies, (5) Optimum Burner Operation, and (6) Nonconventional
Burner Studies.
The section on atomization results is a description of experiments
conducted to determine the spray dropsize distributions produced by
typical oil burner atomizers. These experiments were carried out
using a molten wax technique which allowed collection and sizing of
solidified droplets.
The description of commercial burner studies includes experimentally
measured air velocity vectors, combustion gas velocity vectors, oil
spray patterns, combustion gas composition patterns, flue gas emis-
sions, and combustion chamber wall temperature effects.
The interpretation of commercial burner results includes a detailed
discussion of the rationale behind selection of the versatile burner
design with the rationale being based on the commercial burner studies
and related to consideration of well-stirred versus plug flow combus-
tion processes.
The description of versatile burner studies includes a description
of the reasons for the versatile burner design and a description of
versatile burner geometry variation studies which led to the optimum
burner geometry design criteria.
79
-------
Description of the optimum burner operation involves the fabrication
and operation of two burners, the designs of which were based on the
optimum criteria developed from the versatile burner studies. Included
are comparisons with commercial burners, showing the beneficial effects
of the optimum designs.
The unique concepts section describes experimental results obtained
with a number of burners of very nonconventional design, including
some which are not fired in refractory-lined combustion chambers.
ATOMIZATION MEASUREMENTS
Conical spray nozzles for use with No. 2 distillate fuel oil burners
are available from various manufacturers. Such nozzles vary in flow-
rate, conical spray angle, and spray density patterns. Solid-cone
spray nozzles have a more or less uniform spray mass flux over the
circular spray area, and hollow-cone spray nozzles have high mass
flux near the outer edges of the circular spray pattern. Intermedi-
ate nozzles have lower spray mass flux in some central portions of
the circular area than on the outside, but the difference is not as
extreme as for hoilow-cone sprays. Figure 30 in the Apparatus sec-
tion shows these differences schematically.
Five oil burner spray nozzles were wax flowed, each at three different
pressures, nominally 75, 100, and 125 psig. The experimental data
from these wax flows, in terms of mass median diameter of the collected
frozen wax particles, are presented graphically in Fig. 31. The oil
spray nozzle designations given in Fig. 31 have three terms, e.g.,
0.75-80°C. The first term (0.75) is the rated oil flowrate at 100-
psi pressure drop. The second term (80°) is the full included an-
gle of the spray cone produced. The last term ("C") is an indica-
tion of the nature of the spray pattern, "A" being a hollow cone,
"B" or "R" being a solid cone, and "C" or "W" being a semisolid cone
(see Fig. 30). The nozzles tested were all in the residential-size
range of 0.5- to 1.5-gph rated flow at 100-psig pressure. At the
80
-------
O- DELAVAN 0.75 GPH - 80°-C
D- MONARCH 1.50 GPH - 80°-R
0- DELAVAN 0.75 GPH - 80°-B
A- DELAVAN 1.00 GPH - 80°-C
0- DELAVAN 0.50 GPH - 80°-C
100
I
u
OS
U
1-1
cu
1
70 .
60
80 100 120
INJECTION PRESSURE, PSIG
140
Figure 31. Spray drops!ze data obtained using the method of
frozen wax on various oil burner spray nozzles
81
-------
rated pressure of 100 psig, these residential-size burners typically
produced mass median particle diameters of approximately 75 microns.
Shown in Fig. 32 are mass median droplet size data reported in a Delavan
nozzle brochure. When extrapolated down to the residential nozzle size
range, the value of 75 microns is in agreement with the Delavan curve,
despite the use of a different fluid and a different droplet size ex-
perimental determination technique.
Size distribution curves for the tests conducted at a nominal wax spray
pressure of 100 psig are shown in Fig. 33 through 37. Except for
left-to-right expansion, these curves differ very little in basic shape.
The curves shown are cumulative fraction curves, the derivatives of
which yield the familiar bell curve. All of the distribution curves
shown in Fig. 33 through 37, were obtained by sizing of the collected
particles on a Sharpies micromerograph, a sedimentation-in-air analyzer.
For one case, the particles size analysis was checked by means of stand-
ard sieves, and the comparison shown in Fig. 38 was obtained. It is
noted that the mass median particle diameter determined by the sieving
technique was approximately 10-percent smaller than that obtained by
the micromerograph technique. The reason for the 10-percent discrep-
ancy is not apparent, and its magnitude was not sufficiently large to
provoke a thorough investigation.
COMMERCIAL BURNER STUDIES
To provide a baseline for later development studies, several commer-
cial burners were characterized.
Commercial burners experimentally evaluated to determine pollutant for-
mation and emission characteristics included three burners sized for
residential applications (~0.75 gph) and three larger burners (6 to
12 ghp]. Table 7 presents data pertinent to these burners.
82
-------
CO
z
s
*
OS
120
DELAVAN DATA
a.
i
100
80
I
10 20 30
NOMINAL DISCHARGE RATE, GAL/HR
40
Figure 32. Extrapolation of Delavan data to the
mass median particle diameter obtained
for the Delavan 0.75-gph/80°-C nozzle
83
-------
CO
«
1.0
0.8
8*.
•H O
4J
O
CO
CO M
«
0.4
0.2
o.o
0
20
*'
/
Delavan 0.75-80°-B
100 psi
60
80
100
D , Particle Diameter, Microns
120
140
Figure 33. Spray dropsize distribution for a Delavan 0.75-80°-B nozzle
at 100-psi pressure (raicromerograph analysis)
-------
00
C/1
CO
OJ _P,
r-l O
U
•H e
+> eo
h f.
« P
PH
(0
4-1 00
o eu
s:
^-t 0)
•p +J
U 0)
CO E
£co
3
ca
ca 00
co c
*>:
ca
.0
o"
•
CO
o
•
ON
o
•
f
o
•
I\J
o
•
O
Delavan 0.75-80°-C
100 psi
0
^0
2
OA°
0
x°
h
/
0
/
6c
0
p
)
o
/
8
i
/
f
0
/
1C
/
X>
"*
12
0
U
*0
D i Particle Diameter, Microns
Figure 34« Spray dropsize distribution for a. Delavan 0.75-80°-C nozzle
operating at 100 psi (micromerograph analysis)
-------
CO
CO
(0
to
Pi
-------
00
CO
3
^
I
CO
T( 4)
•P +>
CO
CO 00
0 C
1.0
0.8
0.6
0.1*
0.2
0.0
oocT0!
20
/
60
80
100
D, Particle Diameter, Microns
Monarch 1.50-80°-R
100 psi
120
Figure 36.
Spray dropsize distribution for a Monarch 1.50-80°-R nozzle
at 100-psi pressure [micromerograph analysis}
-------
OO
00
CO
CO Qi
-I Q
y
£
n
n
.3 S3
«*
<0 B
(-1 (D
CO
co
1.0
0.6
0.2
0.0
20
Delavan 0.50-80°-C
100 psi
xO
60
80
100
120
140
D , Particle Diameter, Microns
Figure 37. Spray dropsize distribution for a Delavan 0.50-80°-C nozzle
at 100-psi pressure (micromerograph analysis)
-------
00
(O
n
1.0
0.8
8 0.6
03
§>.
•H V
-P -P
O g
2 I 0.4
0.2
0.0
0
Delavan 0.75-80°-C
75 psi
O Micromerograph
• Sieve
nOO
yd£-
^
s
^ X
X
,
* 0
/
-------
Table 7. COMMERCIAL BURNERS EVALUATED EXPERIMENTALLY
Model
55-J-l
Mite
AFC
Nu-Way "CO"
250 FFD
PHC-34
Manufacturer
ABC
ABC
Union (Beckett)
White-Rodgers
Carlin
Sun Ray Burner
Oil Flowrate,
gph
0.60 to 1.10
1.00 to 1.25
0.50 to 2.50
6.00 to 10.00
5.00 to 12.00
5.00 to 14.00
Type
Nonflame Retention
Flame Retention
Flame Retention
Nonflame Retention
Flame Funnel Retention
Flame Retention
The experimental studies included determination of smoke, unburned hy-
drocarbons, carbon monoxide, and nitric oxide emissions. Typically,
at conditions under which low smoke emissions are obtained, the un-
burned hydrocarbon and carbon monoxide emissions tend to be acceptably
low, while nitric oxide emissions may or may not be low, depending on
the burner design. The crux of the problem, therefore, is achieving
complete fuel combustion (i.e., no smoke, unburned hydrocarbons, or
carbon monoxide in the exhaust), while simultaneously minimizing nitric
oxide formation. Among the experimental results obtained with the
commercial burners, those which have been most enlightening with respect
to nitric oxide minimization are the velocity vector and combustion gas
composition profiles in the combustion zones just downstream of the
end of the burner blast tube.
The results obtained with each of the burners in Table 7 are presented
separately below, after which, the results are discussed and interpre-
tated collectively.
ABC Model 55-J Burner (0.75 gph)
The ABC 55-J is one of the most common types of residential-size No. 2
distillate fuel oil burners in use in the U.S.A. The end of its blast
tube provides a restriction from 4-1/8 to 2 inches, with six peripheral
90
-------
swirler vanes. The oil spray nozzle is located in the center and to-
ward the blast tube end, with electrodes just upstream for purposes
of ignition.
Since the 55-J was accepted as a "standard" oil burner, it was sub-
jected to the most extensive testing of any of the burners used in the
program. Measurements performed with this burner included: (1) cold-
flow air velocity vector measurements in both an 8- and an 11-inch-ID
coaxial cylindrical chamber, (2) cold-flow oil mass flux determinations,
(3) furnace flue gas sampling, (4) combustion gas velocity vector deter-
minations, (5) combustion gas composition pattern measurements, and
(6) mixed combustion gas composition output from an 8-inch-diameter
coaxial cylindrical combustion chamber.
Cold-flow air velocity vector measurements involved determination of
the dynamic head and flow angles a and 3, as described previously in
the Experimental Apparatus section. These measurements were made on
both horizontal and vertical diameters at axial locations 1.5, 3, and
6 inches downstream of the end of the blast tube. The velocity vectors
include the axial, tangential, and radial components of the measured
velocities. Figures 40 and 41 show the experimental data in the hori-
zontal and the vertical planes at the 1-. 5-inch axial location. It is
readily apparent that there is significant nonsymmetry with respect to
the centerline axis (represented by zero radius in Fig. 40 and 41) of
the burner blast tube. The nonsymmetrical characteristics are appar-
ently induced by effects due to the locations of the blower wheel, the
blower inlet, the ignition system, and the static pressure plate, all
of which are relatively close to the exit of the short blast tube. As
previously discussed, the $ angle infers rotation of the flow as it
progresses down the chamber. At the 1.5-inch axial location in the
horizontal plane, the 3 angle reaches a value of approximately ±20
91
-------
-1.
<
5 -1
.»•'
» •
0 -0
• (
>•• •
.5
60 MI-' Ha
..•8*.
20
0
0
•,
* •
•
.5 Radius 1.
Inches
'•.^
0 1
• 5
Figure 40. Air flow parameters at 1.5 inches downstream in
the horizontal plane of a 55-J burner
92
-------
*
x^L
60 MM Hg
20
0
-1.5 -1.0 -0.5 0.5 Radius 1.0 1.5
Inches
Figure 41. Air flow parameters at 1.5 inches downstream in
the vertical plane of a 55-J burner
93
-------
degrees at radii of ±1 inch, while a at the same point is also approxi-
mately 20 degrees. These experimental values can be shown to infer that
the airflow at the point in question is rotating approximately 22 degrees
about the burner axis for each inch of axial travel. It would be ex-
pected, therefore, that should these a and $ angles remain constant, a
90-degree rotation would be achieved in approximately 4 inches of axial
travel. In reality, as shown in Fig. 42 through 45, the B angle de-
creases as the flow continues downstream (primarily due to the cone-
shaped expansion of the flow), and a 90-degree rotation would therefore
require somewhat more than 4 inches of axial travel. A 90-degree rota-
tion of the flow for every few inches of axial travel is realistic. As
shown in Fig. 40 through 45, the nonsymmetry of the flow patterns per-
sists even to the 6-inch downstream axial position.
The figures showing flow angles a and 3 , and dynamic head, AP, accu-
rately describe the flow; however, it is easier to visualize the flow
patterns if they are presented as linear vectors as in Fig. 46 and 47.
The vectors shown in Fig. 46 and 47 represent those components of the
air velocity which are in the plane represented by the paper. In
general, each velocity vector shown in Fig. 42 and 43 has an additional
component (the tangential velocity component) either into or out of
the paper, but that component is not shown since the paper has only
two dimensions.
Assessment was desired of the effects of chamber dimensions on the air
velocity patterns created by the 55-J burner. The measurements carried
out with the 8-inch-ID, coaxial, cylindrical combustion chamber were
repeated with a larger diameter (11 inches) chamber. The results of
the larger chamber dimensions are shown in Fig. 48 and 49 which com-
pare air velocity vectors in the two different diameter chambers. The
difference in chamber dimensions is seen to have a measurable, but not
a significant effect on the flow patterns.
94
-------
-J
-1.
_*-•-<
*+*^
5 -1.
««.*^k=»-<
^
0 -0
.-•-•-•-•-(
.5 C
flP
U) MMH«
--^*"*~'
0
) 0
K""^^,
.5 Radius 1
Inches
K*^
•x..
.0 1.
5
Figure 42. Air flow parameters at 3 inches downstream in the
horizontal plane of a 55-J burner
-------
60
r'\
. -•"
r"
AP
to MM HK
'3r»^
*•*.
1
r*--*--<
P — •--.-
^
-1.5 -1.0 -0.5 0.5 Radius 1.0 1.5
Inches
Figure 43. Air flow parameters at 3 inches downstream in the
vertical plane of a 55-J burner
96
-------
20
15 Degrees
10
-1.5
-1.0
Toward
-0.5
so
IS Degrees
10
-1.5
•l-° ~
o-o-o-Q-0-
0.5 Radius
Inches
1.0
1.5
•10
•20
•25
.
-1
.-.-••-«
•5 -1
i- -•- -•-
.0 -0.
-•- -•- H
5
Af
kO MM KB
20
V""*"
0.
*~ ""^^ ^
5 Radius 1.
Inches
--•--•-
o i.;
-•.
Figure 44. Air flow parameters at 6 inches downstream in the
horizontal plane of a 55-J burner
97
-------
20
15 Degreea
10
-1.5
-1.0
-0.5
Toward
Bottom
-10
1.0
1.5
-20
20
p
15 Degrees
10
- 5
0.5 Radius
Inches
-10
-15
-20
l£5_
— _!
,_-•-•-
-•—•"<
JlO MM Hg
20
0
••- -•- -.
9_
-1.5 -1.0 -0.5 0.5 Radius 1.0 1.5
Inches
Figure 45. Air flow parameters at 6 inches downstream in the
vertical plane of a 55-J burner
98
-------
Geometric Scale: I
Velocity Scale:
PROBABLE
RECIRCULATION
REGION
PROBABLE
RECIRCULATION
REGION
8" Diameter
Chamber
Figure 46. Measured cold-flow air velocity vectors
in the horizontal plane of a 55-J burner
99
-------
\
Geometric Scale:
Velocity Scale:
PROBABLE
^CIRCULATION
REGION
8" Diameter
Chambor
PROBABLE
RECIRCULATION
REGION
Figure 47. Measured cold-flow air velocity vectors
in the vertical plane of a 55-J burner
100
-------
*\\«M;!ii
\\V\\\\\l\\lilili
Velocity Vectors
Represented by
Broken Lines Were
Obtained in ll"-Diam
Chamber
Velocity Scale:
8" Diameter
Chamber
Figure 48. Air velocity vectors for the horizontal plane of
a 55-J burner (comparison of data obtained with
8- and 11-inch-diameter chambers)
101
-------
Geometric Scale:
Velocity Scale:
8" Diameter
Chamber
Z = 6.G"
NOTE; Velocity Vectors
Represented by
Broken Lines Were
Obtained in 11" Diam.
Chamber
.,
BOTTOM
SIDE
TOP
SIDE
Figure 49. Air velocity vectors for the vertical plane of
a 55-J burner (comparison of data obtained with
8- and 11-inch-diameter chambers)
102
-------
Using two different spray nozzles, the local spray mass flux was deter-
mined in the flowfield of the 55-J burner, without ignition of the
spray. These cold-flow measurements were performed as previously
described in the Experimental Apparatus section. The experimental
results are shown in Fig. 50 through 52, overlaid on the previously
presented air flow velocity vectors. For all of the cold-flow tests,
the oil burner air bands were adjusted wide open, which led to the
relatively high air-to-fuel ratios shown in the figures. There is a
very significant lack of symmetry in the spray mass flux, more than
might be expected if the oil spray nozzles were flowed in the absence
of the air blast from the burner. It is concluded, therefore, that
nonsymmetry in the air flow patterns produced by the burner has also
induced nonsymmetry in the oil spray patterns.
Velocity vector measurements were repeated with the burner ignited, so
that combustion gas velocities were being measured. These hot-fire
results are summarized by data presented in Fig. 53 and 54. The veloc-
ities presented in these figures are based on an assumed constant gas
density equivalent to 2300 F temperature and a molecular weight of 29.
Comparison of the combustion gas velocity profiles with the previously
presented cold-flow air velocity profiles show that the essential
features of the patterns, other than magnitude, are not significantly
altered by the presence or absence of combustion.
The 55-J burner was fired into the 8-inch-ID, coaxial, cylindrical
combustion chamber, and experimental techniques previously outlined
were used to sample and chemically analyze local combustion products.
The results of these experiments are presented in Fig. 55 and 56 for
two different overall burner operating stoichiometric ratios. The
parameters presented are local burned gas stoichiometric ratio (based
on oxygen, carbon monoxide, and carbon monoxide, and carbon dioxide
concentrations, and ignoring the presence of liquid or gaseous hydro-
carbons), carbon monoxide, and nitric oxide. Carbon monoxide is ex-
pressed as grams of carbon monoxide per kilogram of burned fuel
103
-------
wf//
Geometric Scale: |
0
Velocity Scale: I •
Fuel Flux
Stoicliiometric Ratio
V
Wf/A
8" Diameter
Chamber
Figure 50. Air velocity vectors and fuel spray mass flux data
for a 55-J burner incorporating a 0.75-80°-C
nozzle (horizontal plane)
104
-------
lOOr
VA 50
Geometric Scale:
Velocity Scale:
150
Wf/A 100
30
Stoichiometric Ratio
Fuel Flux
Wf/A
Ibm/hr-ft
8" Diameter
Chamber
BOTTOM
SIDE
Local
Stoich.
Ratio
A/F
14.49
Figure 51. Air velocity vectors and fuel spray mass flux data
for a 55-J burner incorporating 0.75-80°-C nozzle
(vertical plane)
105
-------
200
Wf//
4 A/F
2 1J^S
0
Geometric Scale:
1"
i
Z = 6.0"
0 30
Velocity Scale: i • • »J '
Ft/Sec
200-
Fuel Flux
~ Stoichionetric Ratio
Wf/A
Fuel Flux
Wf/A
lbm/hr-et"
8" Diameter
Chamber
BOTTOM
SIDE
TOP
SIDE
Figure 52. Air velocity vectors and fuel spray mass flux data
for a 55-J burner incorporating a 1.50-80°-C
nozzle (vertical plane)
106
-------
GEOMETRY SCALE
• = 1.0 IN.
COMBUSTION GAS
VELOCITY SCALE
8" Diameter
Chamber
'/ z = 6.0 IN.
= 3.0 IN.
z = 1.5 IN.
DOWNSTREAM
PUMP
SIDE
MOTOR
SIDE
Figure 53. Combustion gas velocity vectors in the horizontal plane
for the ABC 55-J burner mounted in an 8-inch-diameter,
cylindrical, coaxial chamber
107
-------
Geometric Scale:
Velocity Scale:
30
Ft/Sec
8" Diameter
Chamber
BOTTOM
SIDE
= 3.0"
TOP
SIDE
Figure 54. Combustion gas velocity vectors in the vertical plane
for the ABC 55-J burner mounted in an 8-inch-diameter,
cylindrical, coaxial chamber
108
-------
-i
si
Exhaust Stack
Emission Level
Stoichiometric Ratio
Carbon Monoxide
8
-101
X, Radial Position, Inches
* = Exhaust Stack Emission Levels
Figure 55. Local combustion gas analysis profiles for the ABC
55-J burner at a nominal Stoichiometric ratio of
1.25 with a 0.75-70°A nozzle (horizontal plane)
109
-------
Z = 6.0" Downstream
2.0
C/
1.0
o
I'
*
±
100
-1*0
8
-100
0 o
I
- 100
8
-101
X, Radial Position, Inches
* = Exhaust Stack Emission Levels
Figure 56. Local combustion gas analysis profiles for the ABC
55-J burner at a nominal stoichiometric ratio of
1.50 with a 0.75-70°A nozzle (horizontal plane)
110
-------
(gm NO/kg fuel) in the local gas sample. Nitric oxide is presented
as grams of nitric oxide per kilogram of burned fuel (gm NO/kg fuel)
in the local gas sample. Each of these parameters (as described in
the Apparatus and Data Reduction section of this report) totally
ignores the presence or absence of unburned liquid or gaseous hydro-
carbons. Using previously described experimental techniques, the
measured concentrations of volatile gaseous hydrocarbons in the gas
samples represented by Fig. 55 and 56 were insignificant and, there-
fore, are not presented here. For reference, the pollutant emission
levels in the exhaust stack are shown as asterisks on the ordinates
of Fig. 55 and 56.
Data describing the mixed gases, exiting from the baffled, coaxial,
cylindrical, 8-inch-diameter combustion chamber are given in Fig. 57
through 59. These figures show that carbon monoxide and smoke emis-
sions are low under all conditions for which the 55-J was tested, with
low smoke due most certainly to the absence of fuel-rich stoichiometric
ratios throughout the combustion process, as shown in the composition
pattern data, and with low carbon monoxide emission due to the same
factor as well as the excellent downstream mixing characteristics of
the baffled cylindrical combustion chambers. Data are also shown in
Fig. 57 through 59 for burners other than the 55-J. Those data are
discussed in the sections referring to the specific burners.
Shown in Fig. 60 through 63 are results of flue gas analyses for the
55-J and other burners (discussed later) when fired in the Lennox
forced-air furnace having an upright combustion chamber with the burner
mounted perpendicular to the combustion chamber axis. Carbon monoxide,
unburned hydrocarbons, and nitric oxide emission characteristics do
not show significant differences from the data obtained with the co-
axial chamber. Smoke emissions, however, are considerably higher, due
probably to impingement of liquid fuel on the combustion chamber walls
with decomposition thereon to produce smoke, an event which was con-
siderably less likely with the coaxial chamber than with the upright
perpendicular chamber. Carbon monoxide and unburned hydrocarbons
111
-------
10 _
CQ
UL
tu
O
X
i
0.75-70 -A NOZZLE
ABC 55-J
UNION AFC
— ABC MITE
I
1.0
1.5 2.0 2.5
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
3.0
Figure 57,
Carbon monoxide emissions measured by mixed combustion
gas sampling in the coaxial, 8-inch-diameter
cylindrical combustion chamber
112
-------
2.5.-
2.0
1.5
BQ
I
1.0
W
Q
1-1
X
o
0.5
MITE
55-J
0.75-70 -A NOZZLE
— — — — BACHARACH SMOKE > 1
——__ BACHARACH SMOKE < 1
1.0
1.5 2.0 2.5
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
3.0
Figure 58. Nitric oxide emissions measured by mixed combustion
gas sampling in the coaxial, 8-inch-diameter
combustion chamber
113
-------
10 —
CO
I
VI
X
0
0.75-70 -A OIL NOZZLE
ABC MITE
UNION AFC
ABC 55-J
1.0 1.5 2.0
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
2.5
Figure 59. Smoke emissions measured by mixed combustion
gas sampling, coaxial 8-inch-diameter
combustion chamber
114
-------
Ib
E
g
« 10
"3
3
U.
O
1
*
«"
•o
X
o
s
i
B
1 5
efl 3
u
0
-
™
I
1
I
1
I
1
1
;
i
/
/
'
,' I
/ i
• i
• i
, j
1 4 1
/
1 /
/ /
1 1 •
1 /
i i
i i
I 1
L U
0.75-70°-A Nozzle
ABC 55-J
Union AFC
ABC Mite
1.0
1.5 2.0 2.5
Stoichiometric Ratio, (A/F)/14.49
3.0
Figure 60. Carbon monoxide emissions measured by Lennox
furnace flue gas sampling
115
-------
2.5-
2.0
g
a.
u.
o
1.5
g i.o
t-l
0.5
0.75-70 -A OIL NOZZLE
=BACHARACH SMOKE < 1
— BACHARACH SMOKE > 1
ABC 55-J
UNION AFC
ABC MITE
I
1.0
1.5 2.0 2.5
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
3.0
Figure 61. Nitric oxide concentrations measured in the
Lennox furnace flue gas samples
116
-------
o.u
7.0
1
I 6.0
H
J
° 5.0
B
•§
to
* 4.0
8
o
0
CO
u
I 3.0
w
•a
v
g
1 2.0
£
1.0
0
,« 0.75-702A Oil Nozzle
'i ... ,
/I ABC 55^J
/ Union AFC
/ ABC Mite
_ /
/ 1
/ 1
1
1
1
i
1
1
1
1
1
•
i
1
1
1
1- B
1 /
1 ' /
1 i '
1 ' /
1 / /
> //
1.0
1.5
2.0
2.5
Stoichiometric Ratio,
3-0
Figure 62. Total hydrocarbon emissions measured by Lennox
furnace flue gas sampling
117
-------
3
§ 5
0
1.0
ABC 55-J
UNION AFC
0.75-70 -A OIL NOZZLE
1.5 2.0
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
Figure 63. Smoke emissions measured by Lennox
furnace flue gas sampling
2.5
118
-------
might also result by the same mechanism; however, their nature is such
that they can be eliminated downstream by oxidation, and their absence
in the flue gas is therefore not surprising, assuming that only moder-
ate amounts of fuel impinge on the upright combustion chamber walls.
Impingement of only a moderate amount of fuel on the upright combustion
chamber walls is consistent with the chamber dimensions and the combus-
tion gas flow patterns presented previously as a function of distance
from the blast tube end (assuming that similar patterns might result
when the 55-J is fired into the upright chamber).
ABC Mite Burner (1.0 gph)
The ABC Mite burner is unique because of the relatively small (3-1/6
inch) OD of its blast tube. It has a 2-1/2-inch-ID choke plate in the
end of the blast tube, and a 2-inch-diameter air spinner with windmill-
like vanes suspended in the center of the choke plate. The oil spray
nozzle is centered within the air spinner.
Cold-flow air flow patterns produced by the Mite burner were relatively
complicated, as shown by the data given in Fig. 64 through 66. At the
1.5-inch location, the 3 angle parameter, which indicates air swirl, is
most different from zero at the locations 0.5 to about 1.2 inches radius
from the center of the burner. At the 3- and 6-inch downstream locations,
the air flow parameters (Fig. 65 and 66) indicate that the flow patterns
are less complicated and, in particular, the localized swirl at the 0.5-
to 1.2-inch radial locations is considerably less pronounced.
Two-dimensional representations of the air flow velocity vectors are
given in Fig. 67 and 68. The swirl (tangential velocity) component
cannot be shown in these figures, but the high degree of flow irregu-
larity at the 1.5-inch coaxial location is nonetheless still very evi-
dent, with less irregularity at the more downstream locations.
119
-------
-l.Oi
-0.5
Radius, Inches
Figure 64.
Air flow parameters at 1.5 inches downstream
of the Mite burner in an 8-inch-diameter,
cylindrical, coaxial chamber (zero oil
flow, vertical plane)
120
-------
0.5 1.0 1.5
Radius. Inches
AP, MM Kg, Dynamic Head
1.5
Figure 65. Air flow parameters at 3 inches downstream of the
Mite burner installed in an 8-inch-diameter
cylindrical, coaxial chamber (zero oil flow,
vertical plane)
121
-------
0.5 l.U 1.5
c Radus, Inches
0.5 1.0 1.5
Radius, Inches
A P. MM Hg, Dynamic Head
Figure 66. Air flow parameters at 6 inches downstream of the
Mite burner in an 8-inch-diameter cylindrical,
coaxial chamber (zero oil flow, vertical plane)
122
-------
100
Wf/A
Geometric Scale:
1"
I
_A/F_
11*. 49
Jo
Z = 6.0"
0 30
Velocity Scale: | . . _|
Ft/Sec
100 -
Fuel Flux
— Stoichiometric Ratio
Fuel Flux 10°
VA
Ibm/hr-ft^
8" Diameter
Chamber
5 Local
Stoich.
Ratio
A/F
T4~4T
PUMP
SIDE
MOTOR
SIDE
Figure 67. Air flow and oil spray flux patterns in the horizontal
plane of the Mite burner mounted in an 8-inch-diameter
cylindrical, coaxial chamber (Delavan 1-80°-C nozzle)
123
-------
100
-^mMlrnr/TTT^.
Geometric Scale: I i_
Velocity Scale: L
x tJ
Z = 6.0"
30
Fuel Flux
Stoichiometric Ratio
100
Wf/A
0L
Fuel Flux
Wf/A 100
lbm/hr-ft2
Local
Stoich.
Ratio
« A/F
0 14.49
8" Diameter
Chamber
BOTTOM
SIDE
.
y
TOP
SIDE
Figure 68. Air flow and oil spray flux patterns in the vertical
plane of the Mite burner, mounted in an 8-inch-
diameter, cylindrical, coaxial chamber
(Delavan 1-80°-C nozzle)
124
-------
Also shown in Fig. 67 and 68 are cold-flow (nonignition) oil spray
flux patterns measured by techniques previously described, with the
air band wide open. Note that the oil spray flux appears to be some-
what rarified at the 0.5- to 1-inch radial positions, where it has
been noted there exists a high local rate of swirl. These low spray
mass fluxes may be a combined result of the spray nozzle design, flow
angles, and centrifugal forces, caused by the air swirl, slinging the
larger-size oil droplets out of the zone of rotating gases. This same
effect is noted in both the horizontal and the vertical planes (Fig.
67 and 68, respectively).
Combustion gas velocity vectors, measured with the burner ignited, are
shown in Fig. 69 and 70. Except for magnitude, there is a pronounced
similarity between the combustion gas velocity vectors and the cold-
flow air velocity vectors. Irregular patterns of the combustion gas
velocity vector profiles do propagate further downstream than the
analogous irregularities in the air velocity vector profiles.
Local gas sampling for determination of pollutant concentrations was
performed for the Mite burner fired into the 8-inch-ID coaxial, cylin-
drical combustion chamber. The results of the sample gas analyses are
shown in Fig. 71 through 73 for overall burner operating stoichiometric
ratios of 1.25, 1.50, and 1.80. The most interesting facet of these
results is the tendency for very high nitric oxide concentrations,
near unity stoichiometric ratios, and minima in carbon monoxide con-
centrations at the 0.5- to 1.2-inch radial locations where it was
previously noted that there exists a high degree of local gas swirl.
At these locations, the closeness to unity local stoichiometric ratio
and the magnitude of the nitric oxide concentration peaks are most
pronounced for the axial locations nearest the blast tube (1.5-inch
axial rotation) and for the overall stoichiometric ratios nearest
unity. The high temperatures which accompany near unity local stoi-
chiometric ratio and the poor chances for heat loss due to remoteness
from the chamber walls undoubtedly are major factors leading to the
125
-------
GEOMETRY SCALE
COMBUSTION GAS
VELOCITY SCALE
8" Diameter
Chamber
= 1.0 IN.
= 30 FT/SEC
z=6.0 IN.
0 IN.
.2=1.5 IN.
DOWNSTREAM
PUMP
SIDE
MOTOR
SIDE
Figure o9. Combustion gas velocity vectors in the horizontal plane
for the ABC Mite burner mounted in an 8-inch-diameter,
cylindrical, coaxial chamber
126
-------
Geometric Scale: i
Velocity Scale: | ,,
8" Diameter
Chamber
BOTTOM
SIDE
TOP
SIDE
Figure 70. Combustion gas velocity vectors in the vertical plane
for the ABC Mite burner mounted in an 8-inch-diameter,
cylindrical, coaxial chamber
127
-------
_ 200
3 r
-- 100.
- 200
!
~-~~~^- loo g
gS
I
i
1.5" Downstream ^s
i
8
i
_ 100
8
-101
X, Radial Position, Inches
Exhaust Stack Emission Levels
Motor Side
Figure 71. Combustion gas analysis profiles for the ABC Mite burner
at a nominal stoichiometric ratio of 1.25 with a
0.75-70°-A nozzle
128
-------
-1200
- 100
-.200
- 100.
+>g. s|e 1.5" OownEtrean
I
\
8
I
I
5
8
— 200
i
- 100.
8
-3
-101
X, Radial Position, Inches
Exhaust Stack Emission Levels
Figure 72. Combustion gas analysis profiles for the ABC Mite burner
at a nominal stoichiometric ratio of 1.50 with a
0.75-70°-A nozzle
129
-------
-1200
- 100
- 100
*o g
n
!
8
200
^5
I
- 100
8
-10 1
X, Radial Position, Inches
* = Exhaust Stack Emission Levels
Motor Side
Figure 73. Combustion gas analysis profiles for the ABC Mite burner
at a nominal stoichiometric ratio of 1.80 with a
0.75-70°-A nozzle
130
-------
high local nitric oxide concentrations. Comparison of the nitric oxide
at the 1.5-, 3-, and 6-inch axial locations indicates a monotonic de-
crease in the magnitude of the nitric oxide peaks which must be an in-
dication of mixing of the local gases with surrounding gases, rotation
due to swirl transporting the nitric oxide-laden gases to a different
angular location and/or chemical change of the nitric oxide. Of these
three possibilities, the first two seem most likely. Since composition
measurements were made only in the horizontal plane of the burner, it
is not possible to state whether the new gases replacing the old, due
to swirl, were originally high in nitric oxide, but it seems very likely
that they were, assuming symmetry. Degradation of the nitric oxide
peaks is, therefore, most probably due to mixing of the high nitric
oxide concentration gases with surrounding gases having lower nitric
oxide concentrations. This seems highly probable in view of the fact
that gas velocities (Fig. 69 and 70) in the annular-shaped, high nitric
oxide zones are only 1/4 to 1/3 the velocities of the surrounding gases,
a fact that tends to enhance mixing and which indicates that the volume
flux of the high nitric oxide concentration gases is considerably lower
than the volume flux of the surrounding low nitric oxide concentration
gases with which it is apt to mix.
As the overall stoichiometric ratio is raised from 1.25 to 1.50, to
1.80, the nitric oxide peaks become less pronounced, apparently because
the local stoichiometric ratio becomes further removed from unity.
Carbon monoxide concentrations start low at the 1.5-inch axial loca-
tions, become higher at the 3-inch axial location, then become dissi-
pated by the time the gases reach the 6-inch axial location. This
behavior is as expected for this combustion intermediate reaction
product.
The nitric oxide data presented in Fig. 71 through 73 seem to have a
tendency toward minima when the local stoichiometric ratio is in the
neighborhood of 1.80, or 80-percent excess air. This general trend
also was noted with other burners.
131
-------
Compositions of the mixed combustion gas from the coaxial cylindrical
combustion chamber were previously presented (Fig. 57 through 59)
together with the similar data for the 55-J burner. From the coaxial
cylindrical combustor, the mixed ABC Mite combustion gases showed low
carbon monoxide (probably more due to the large combustor volume than
the particular burner), relatively high nitric oxide, and relatively
high smoke. The high nitric oxide was undoubtedly provoked, at least
partly, by the previously described combustion conditions in the 0.5-
to 1.2-inch radius annular locations near the blast tube exit. The
relatively high smoke level was a result of the wide variations in
local stoichiometric ratio which caused some localized spots having
fuel-rich burned gas stoichiometric ratios.
The furnace flue gas composition obtained with the Mite burner (Fig.
60 through 63) are indicative of the occurrence of poor combustion due
to combustion stability problems of the Mite burner/Lennox furnace
combinations, as operated for these tests. The result was high carbon
monoxide and smoke coupled with low nitric oxide emissions, the funda-
mental symptoms of poor combustion wherein large variations resulting
in fuel-rich local stoichiometric ratios occur.
Union Model AFC Burner (0.75 gph)
The third residential-size burner tested was the Union Model AFC burner.
The unique construction feature of this burner involves a flame-
retention head with a central sheet-metal flame cone which is moveable.
Movement of the flame cone causes the air flow to be more or less
restricted at the blast tube exit, depending on the direction of the
movement. The restriction of the flame cone forces air passing out
the blast tube exit to flow along peripheral swirlers located on the
choke plate of the b'last tube exit.
Air velocity vectors for the Model AFC burner, obtained under cold-
flow conditions, are presented in Fig. 74 and 75. The nature of the
burner design is such that the air flow is spread out over a relatively
132
-------
Geometrical Scale: L
0
Velocity Scale: L
1"
J
30
J
Local Fuel
Flux, Wf/A
lbm/hr-ft2
-* Z = 1.5'jJ
»-»_Z = 0.75"
8" Diameter
Chamber
PUMP
SIDE
MOTOR
SIDE
Figure 74. Air flow and oil spray flux patterns in the horizontal
plane of the model AFC burner mounted in an 8-inch-
diameter cylindrical, coaxial combustion chamber
(Delavan 0.75-70°-C nozzle)
133
-------
Geometrical Scale: I
1"
I
Local fuel
Flux, Wf/A
lbm/hr-Ft2 100
L
8" Diameter
Chamber
BOTTOM
SIDE
0 30
Velocity Scale: I i i ^J
Ft/Sec
Local
5 Stoich
Ratio
A/F
14.49
///in\\\\//l
« Z = 1.5"-*J
Z = 0.75"
TOP
SITE
Figure 75. Air flow and oil spray flux patterns in the vertical
plane of the model AFC burner mounted in an 8-inch-
diameter, cylindrical, coaxial chamber (Delavan
0.75-70°-C nozzle)
134
-------
large area, leading to low local air velocities. Note that for this
burner, air velocity measurements were made at 0.75- and 1.50-inch
axial positions only, since gas velocities were much lower at the more
downstream positions, making the measurements more difficult. The
design of the flame-retention head for this burner is such that air
from the blast tube is directed radially inward to a center impinge-
ment point very near the blast tube exit. After the air impinges upon
itself, it reflects in a relatively outward direction, ultimately lead-
ing to the well-dispersed flow observed at the 1.5-inch axial location.
Also shown in Fig. 74 and 75 are the local spray mass fluxes determined
from cold-flow experiments with the air band wide open. These spray
mass fluxes have minima similar to those observed when the same spray
nozzle was tested with the ABC Mite burner. The existence of the min-
ima with the Model AFC burner, which has no highly local swirl such
as the ABC Mite, indicates that the minima may be the result of the
spray nozzle characteristics rather than those of the burner air flow.
Local combustion zone gas samples were obtained with the Model AFC
burner mounted in the 8-inch-diameter coaxial cylindrical combustion
chamber. The data obtained are presented in Fig. 76 and 77. The
relatively uniform mixing characteristics of this burner are evident
from these composition profiles. The most significant factor is the
effect of overall stoichiometric ratio on the nitric oxide concentra-
tions. The nitric oxide concentrations are lower at the 1.50 stoichio-
metric ratio than they are with the 1.25 stoichiometric ratio, con-
sistent with previous trends of nitric oxide versus local stoichiometric
ratio indicated minimum nitric oxide at stoichiometric ratios near
1.8. Carbon monoxide levels are lowest for the more air-rich operating
condition.
Compositions of Model AFC burner mixed exhaust gases from the coaxial
cylindrical combustion chamber and from the Lennox furnace were pre-
sented earlier, along with data from the 55-J burner in Fig.57 through
63. Those results are consistent with the good mixing behavior of the
135
-------
3
O *t
ft 1
0
Stoichiometric Ratio
Carbon Monoxide
Nitric Oxide
[_____ 1
* 1.0
o£
w
Z = 6.0" Downstream
'
1
8
2 » 3.0"
• 100'
-1 0 Jl
X Rndinl Position, Inches
+2
-110 =
* = Exhaust Stack Emission Levels
Figure 76. Combustion gas analysis profiles for the model AFC burner
at a nominal Stoichiometric ratio of 1.25 with a
0.75-70°-A nozzle
136
-------
Stoichiometric Ratio
Carbon Monoxide
Nitric Oxide
*
:t
1.0
OS.
»
C/J
Z = 6.0" Downstream
100
n
I
1*0 g
II
-1100
-3 -P -1 0 1 2
X Raflial Position, Inches
* = Exhaust Stack Emission Levels
Figure 77. Combustion gas analysis profiles for the model AFC burner
at a nominal Stoichiometric ratio of 1.50 with a
0.75-70°-A nozzle
137
-------
Model AFC burner and the expected influence of uniform local stoichio-
metric ratio on pollutant emissions. With respect to smoke, hydrocar-
bon, and carbon monoxide emissions, the Model AFC burner had a rela-
tively wide acceptable overall stoichiometric ratio operating range
(from stoichiometric ratio ~1.1 to ~2.2) which was free of signifi-
cant smoke and CO emissions. The uniform mixing of the Model AFC burner
has, therefore, narrowed the spread of local stoichiometric ratios so
that, for example, it is possible to operate at an overall stoichiometric
ratio relatively close to unity, without having too much local flow at
below unity stoichiometric ratios.
Nu-Way Model CO Burner (6.0 gph)
The Nu-Way Model CO burner is a larger than residential-size burner
capable of operating in the 6- to 10-gph oil flowrate range. For the
experiments conducted, this burner was fired at 6 gph. Construction
of the Nu-Way is similar to the 55-J, except it is considerably larger
in size, and utilizes dual, side-by-side 3-gph spray nozzles.
Air flow velocity vectors were determined for the Nu-Way burner, as
shown in Fig. 78 and 79. In general, the air velocities with this
burner were relatively high (~60 ft/sec) compared with velocities
observed in the residential-size burners (~30 ft/sec). Note that
the velocity vector scale is 60 ft/sec/in, in Fig. 78 and 79 compared to
30 ft/sec/in, for previous presentations. The higher air flow pressure
drops coincident with the higher gas velocities lead to a relatively uni-
form distribution of air velocities near the blast tube exit of the Nu-
Way burner.
Fuel spray mass fluxes, experimentally determined under cold-flow
conditions, also are shown in Fig. 78 and 79. The spray mass fluxes
are not very uniform at the 1.5-inch axial location, but they become
relatively uniform at the 3- and 6-inch axial locations, except near
the outer edges of the flow patterns.
138
-------
1000
Wf/A
Geometric Scale:
Velocity Scale: 1
1000 [_
W,/A p
f OL
4000
Fuel Flux
Wf/A
Ibm/hr-ft
2000 .
11" Diameter
Chamber
PUMP
SIDE
Ft/Sec
Fuel Flux
Stoichiometric Ratio
A/F
14.49
Local
Stoich,
Ratio
A/F
14.49
MOTOR
SIDE
Figure 78. Air flow and oil spray flux patterns in the horizontal
plane of the Nu-Way burner mounted in an 11-inch-
diameter, cylindrical, coaxial chamber (two 3-gph
Delavan nozzles)
139
-------
2000
Wf//
Geometric Scale: I
Velocity Scale: I
Ft/Sec
Wf/A
Fuel Flux
W£/A
lbm/hr-ft2
11" Diameter
Chamber
BOTTOM
SIDE
Fuel Flux
Stoichiometric Ratio
2000 -
0
4000
2000
Figure 79. Air flow and oil spray flux patterns in the vertical plane
of the Nu-Way burner mounted in an 11-inch-diameter,
cylindrical, coaxial chamber (two 3-gph Delavan nozzles)
140
-------
Combustion gas velocity vectors were also measured with the Nu-Way
burner; however, since the flame front was displaced approximately
4 inches away from the blast tube exit (due to high air velocities),
combustion gas velocity values are reported only at the 6-inch axial
location. The combustion gas velocity profiles (Fig. 80 and 81)
appear to be less uniform than the air velocity patterns at the same
location, most likely due to the disturbance-amplifying pressure
gradient classically present at distinct flame fronts.
The Nu-Way burner was fired in the 30-inch-diameter coaxial combustor,
with the combustion chamber configured for cold (water-cooled) wall and
hot (refractory-lined) wall. The results for the cold-wall and the hot-
wall tests are shown in Fig. 82 and 83. As shown in Fig. 82, the burner
was also fired in an 11-inch-diameter coaxial combustion chamber to
assess the effects of chamber size on nitric oxide emissions. The
nitric oxide emission results shown in Fig. 82 provide insight into
the effects of both chamber size and chamber wall temperature.
Lowering the wall temperature by changing from a refractory liner
(2500 to 3000 F wall temperature) to a water-cooled wall has a very
definite beneficial effect on nitric oxide emissions. This effect
is undoubtedly due to the lowering of the flame temperature in the
combustion chamber by: (1) convective heat losses from recirculating
combustion gases to the cooled wall, and (2) radiative heat loss from
the core of the flame zone to the cooled walls.
Decreasing the chamber diameter appears to have detrimental effect on
nitric oxide formation. This result may be due to combustion pattern
changes in the core of the flow, but it is probably more predominantly
a thermal effect resulting from: (1) reduced opportunity for combus-
tion gas recirculation and (2) reduced total heat loss through the
smaller area walls of the 18- and 11-inch combustion chambers. The
141
-------
Z = 6.0"
Geometrical Scale: I
1"
J
Approximate Location
of Flame Front
30
Velocity Scale:
11" Diameter
Chamber
PUMP
SIDE
Ft/Sec
MOTOR
SIDE
Figure 80. Combustion gas velocity vectors in the horizontal plane
for the Nu-Way burner mounted in an 11-inch-diameter,
cylindrical, coaxial chamber
142
-------
Z = 6.0"
Approximate Location
of Flame Front
Geometrical Scale:
1"
J
0 30
Velocity Scale: | . . i
Ft/Sec
11" Diameter
Chamber
BOTTOM
SIDE
TOP
SIDE
Figure 81. Combustion gas velocity vectors in the vertical plane
for the Nu-Way burner mounted in an 11-inch-diameter,
cylindrical, coaxial chamber
143
-------
2.5
18-IN. DIAMETER
COMBUSTOR
2.0
u.
o
g
W
g
(HOT WALL)
ll-IN. DIAMETER
COMBUSTOR(SMOKE MEASURE-
MENTS NOT TAKEN)
0.5
30-IN. DIAMETER COMBUSTOR
(COLD WALL)
6 GPH NU-WAY BURNER
HOT WALL EXPERIMENTS
BACHARACH SMOKE < 1.0
BACHARACH SMOKE 2-1.0
1.00 1.10 1.20 1.30 1.40
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
1.50
Figure 82. Effect of combustion chamber size
on nitric oxide emissions
144
-------
HOT WALL
COLD WALL
1.0
1.04
1.08
1.12
1.16
1.20 1.24
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
Figure 83. Smoke content of the exhaust gases for the 6-gph
Nu-Way burner operating in the 30-inch-diameter,
coaxial combustion chamber under cold- and hot-
wall conditions
145
-------
selection of a chamber diameter intermediate between the 30-inch cham-
ber appears to be a reasonable compromise between nitric oxide emis-
sions and combustion chamber capital cost. According to usual indus-
trial practice, the Nu-Wax burner, at 6.0 gph, would normally be fired
in a combustion chamber intermediate in volume between the 18-inch-
diameter chamber (nominally sized for 3 gph) and the 30-inch-diameter
chamber (nominally sized for 12 gph). The fact that (according to
Fig. 82) a compromise between chamber capital costs and nitric oxide
emissions also leads to selection of a chamber diameter intermediate
between 18 and 30 inches suggests that typical burner manufacturer
design criteria are not unreasonable with respect to nitric oxide
emissions. Burner manufacturers claim their chamber size criterion
is that the chamber should be sufficiently large to have the entire
luminous zone of combustion contained within the chamber.
Shown in Fig. 83 are the Nu-Way smoke emissions as a function of stoi-
chiometric ratio and chamber wall temperature. The cooler wall tem-
peratures lead to higher smoke emissions. This is apparently due to
a greater tendency for the cool walls to quench the combustion before
its completion. Note that if the usual procedure of setting the burner
at number one smoke were followed, the nitric oxide emissions would be
nominally the same for both the hot-wall and the cold-wall conditions.
For this burner, carbon monoxide emissions were generally negligible
when the smoke number was less than one.
Local combustion gas samples also were withdrawn and analyzed during
tests with the Nu-Way burner in the 30-inch chamber. The stoichiomet-
ric ratio, carbon monoxide, and nitric oxide profiles obtained during
the hot-wall tests are shown in Fig. 84 and 85 as functions of radial
position (r), and axial position (z) relative to the end of the blast
tube centerline. The same data are replotted on a reduced scale in
Fig. 86, and compared with similar data from the same burner operating
under cold-wall conditions. The cold-wall/hot-wall comparisons shown
146
-------
s r
-.200
o> a
IT w
30 Inch Diameter Hot Wall Combustor
STOICHIOMETRIC RATIO
CARBON MONOXIDE
NITRIC OXIDE
5 10
r. Radial Position. Inches
0 5 10
r, Radial Position, Inches
* • Exhaust Stack Emission Levels
Figure 84. Hot-wall, local combustion gas analysis profiles for the
6-gph Nu-Way burner at nominal stoichiometric ratio of
1.16, with twin 3-60°-B oil nozzles
147
-------
o> 4 r
". S
*•* o£
/-«ta
w J
2 LU Z
U. D
1-4 U.
< o
it '
«k
-
"~^x
30 Inch Diameter Hot Wall Combustor
ST^ICHIOMETRIC RATIO
CARBON MONOXIDE
NITRIC OXIDE
2 = 15.0 In.
••
M
\
U Ul
f* X 1
u o
l-l I-H
gm
H
|S o
-
»k ^^
\ ^^^^
^
" — ^^
1
r
•
o
111
150 |
1
UJ
£
100 °
1
•>^
a
A
(U
Q
50 g
|
•z
01
-J
111 -J
3 U
u. •=>
•
u iu
t-H O
g rs
5 10
r. Radial Position, Inches
z = 12.0 In.
14
150
100 o
50 S
X
0
5 10
r, Radial Position, Inches
* = Exhaust Stack Emission Level
14
Figure 85. Hot-wall, local combustion gas analysis profiles for the
6-gph Nu-Way burner at a nominal stoichiometric ratio of
1.16, with twin 3-60°-B nozzles
148
-------
4 Overall Stoichiometric Ratio = 1.16 ~ 30-Inch Diameter ChamberI
5- 200-, , .. . , . ... 5_
SR.
2 = 6
15
CO.
Cold
Hot Wall
z = 6
IS
NO
0
s_
,
SR.
0
200-
z = 9
CO.
-"aj-- m---i^*-
0
5
z = 9
XNO-
•
-
_
z = 9
A
' V
V , i i. I,
15
5
;
SR
0
200-
z"12 CO.
" ~l 1 1 •! 1 II • •
n
z = 12
^MPW
VV
*cv
15
5
*
SR.
0
z = 15
15
200,
z = 15
S-,
NO
z - 15
15
Figure 86. Comparison plots of local combustion zone sampled Stoi-
chiometric ratio (SR, AIR/14.49 FUEL), carbon monoxide
(gm CO/kg of fuel burned), and nitric oxide (gm NO/kg
of fuel burned), as a function of radial location
(r = inches) and axial location (z = inches)
(Nu-Way burner, 5.5 gph)
149
-------
in Fig. 86 suggest several interesting observations. Comparison of
the stoichiometric ratio profiles suggests that the oil is vaporized
and burned slightly earlier in the hot-wall case, as might be expected
under conditions of higher temperature. Comparison of the carbon mon-
oxide profiles shows no significant differences except at the 6-inch
axial location where the hot-wall profile is more widespread, indicat-
ing active combustion processes at larger radii, perhaps due to earlier
ignition from hotter recirculating gases. The nitric oxide content
has a tendency to peak at radii of 2 to 3 inches, perhaps due to the
competing effects of increased nitric oxide due to approaching unity
stoichiometric ratio at the larger radii and decreased nitric oxide
due to intermixing of the combustion gases with the cooled recirculated
gases at the larger radii.
Comparison of the nitric oxide profiles shown in Fig. 86 for hot wall
and cold wall indicates the hot-wall nitric oxide is significantly
higher than the cold-wall nitric oxide only in the region very near
the end of the blast tube (i.e., at z = 6 inches). This is reasonable
because, when the burner is operating at steady state, the ignition
front is maintained by intermixing of hot recirculated gas with the
injected air/fuel-spray mixture; and, in the hot-wall case, these
intermixed gases are hotter and, therefore, more prone to induce high
nitric oxide. The nitric oxide content of the recirculating gases in
both hot- and cold-wall cases is about 25 percent lower than the 1.3
and 1.0 gm NO/kg fuel observed in the mixed flue gases, from the two
wall conditions, indicating that either: (1) the recirculating gases
do not include gas from nitric oxide-rich core of the flame, (2) some
additional nitric oxide is formed by thermal processes as the gases
pass through the remaining 6 to 8 feet of the baffled exhaust gas mixing
section of the combustor, or (3) a combination of these effects. It
would seem from inspection of the data that the most likely of these
three alternate reasons is (1).
150
-------
As shown in Fig. 87, composition profiles also were measured when the
Nu-Way burner was fired in the 11-inch-diameter hot-wall chamber. Also
shown in Fig. 87 are comparative 30-inch-diameter cold-wall data. Both
sets of data were taken at 9 inches downstream of the blast tube exit.
Surprisingly, at this axial location, the combustion appears to be more
nearly complete in the 11-inch chamber, so that the composition profile
at the 9-inch axial location in the 11-inch diameter chamber more nearly
resembles the composition profiles (Fig. 86) at the 12-inch axial loca-
tion in the 30-inch-diameter chamber. This may be due to the flame
being held closer to the blast tube in the 11-inch-diameter chamber
as a result of increased radiant heat exchange from the walls of the
smaller diameter chamber.
Nu-Way Burner Radiation Measurements. The 6-gph Nu-Way burner was
fired in the 11-inch-diameter cylindrical combustion chamber for radi-
ant energy measurements. A Land Instruments Inc., 2ir steradians radi-
ometer was used to measure the radiant energy flux profile across
radial traverses at five axial locations of Z = 1.5, 3, 6, 9, and 12
inches downstream of the burner head. The radiant energy flux pro-
files are presented in Fig. 88 and 89, showing data for nominal stoi-
chiometric ratios of 1.25 and 1.50, respectively. The Land radiometer
measures radiant energy flux by means of the temperature rise of a
blackbody contained in a highly reflective cavity with a small hole for
admission of the radiant energy. The radiometer is calibrated and
operated with a positive gaseous nitrogen purge to prevent combustion
contaminants from entering the blackbody cavity. However, oil spray
from the Nu-Way oil burner proved to be capable of penetrating the
protection provided by the gaseous purge. Accumulation of small amounts
of fuel oil within the sensor head cavity caused deterioration of the
black, heat-absorbent coating. Therefore, the data presented in
Fig. 88 and 89 should be used for qualitative evaluation only, as
the calibration curve probably shifted with the deterioration of the
coating. Radiometric probing was initiated at the 12-inch location
which, with the least amount of coating deterioration, should be the
151
-------
en
to
METRIC RATIO
— — — CARBON MONOXIDE
uT-TUTr- nvTnt:
z o 9.0 In. z = 9.0 In.
11-In. Diameter 30- In. Diameter
So 4
• UJ
rH 2
13 U4
*^
HH {i,
d-I
P > 2
Cc ^^
*
HIOMETRIC
RIG OXIDE
i-i
U H
H H
O X
85 0
p. Combustion Chamber , , Combustion Chamber n
Hot-Wall
-
_
.
_ «•— — •
*
*
^ ^ ^^^^^^^^^
/
i^r
Cold-Wall w
\ --N
* _ k
/ \ \
* \y x
^^^^^^^ \
^^^^^ V^ y
V-
^"^*
1
150 g
CO
1
u.
u.
100 °
J>
JK
DO
s
1— 1
so g
z
§
0 §
-S
+S
r, Radial Position, Inches
Figure 87. Comparison of combustion gas composition profiles of the
6-gph Nu-Way Model CO burner in hot- and cold-wall
enclosures of different diameters
-------
-5
Figure 88.
-l* -3 -2
X, Radial Position, Inches
Radiant energy profiles for the 6-gph Nu-Way
burner at a nominal stoichiometric ratio of
1'.25 (at various distances downstream)
153
-------
-5
-3 -2
X, Radial Position, Inches
-1
Figure 89. Radiant energy profiles for the 6-gph Nu-Way
burner at a nominal stoichiometric ratio of
1.50 (at various distances downstream)
154
-------
most accurate for any quantitative evaluation. At the 1.5-, 3-, and
9-inch axial locations, a rapid decrease in indicated radiant energy
flux was observed between the radial positions of 3 to 4 inches from
the burner axis. This drop in indicated flux may be due either to a
change in the flame-holding characteristics of the probe itself (if
any) or, more likely, to the presence and evaporation of oil within
the blackbody cavity. The low, flat curves at 6 and 9 inches also
may have been influenced strongly by oil accumulation in the probe.
The data presented in Fig. 89 indicate that an average radiant flux
2
of 10 to 15 watts/cm might be expected at any typical location. It
is of interest to determine how important this energy flux might be
relative to the droplet vaporization and combustion processes. The
flux, as measured, is received from 2ir steradians (a hemisphere) just
as the radiant flux would be received by a differential segment of
droplet surface area. The radiant energy flux can therefore be com-
pared with the energy flux required per unit droplet surface area to
vaporize a typical droplet at a rate characteristic of burning. The
total flux for droplet vaporization is received by a combination of
forced convection and radiation. For a droplet burning in stagnant
air, the burning rate can be described by the burning rate parameter
k' as follows:
k' = dt PL TTD
where
k1 = burning rate constant, cm /sec
D = droplet diameter, centimeter
t = time, seconds
m = droplet mass, gram
p. = droplet density, gin/cm
L
155
-------
The above expression is easily used to determine the necessary energy
flux to sustain droplet vaporization at a rate consistent with the
burning rate constant k1:
- (dm/dt) AHv k1 QL AHy
A = JJJJ2 = 4~D
where
Q/A = thermal energy flux per unit droplet surface area,
cal/cnr-sec
AH = latent heat of vaporization, cal/gm
For a typical hydrocarbon droplet burning in stagnant air, k1 =* 0.01
2
cm /sec (see, e.g., Goldsmith, M. and S. S. Penner, "On the Burning
of Single Drops of Fuel in an Oxidizing Atmosphere," Jet Propulsion.
V. 24, pp 245-251, 1954), PL sO.8 and AHy s 100 cal/gm. Typical
droplet size in the burner tested is about 0.01 centimeter and, for
these parameters, the above equation yields a required energy flux
2 2
for vaporization of 20 cal/cm -sec or about 83 watt/cm . The observed
2
radiant energy flux of-15 watt/cm is about 18 percent of the 83 watt/
2
cm required for vaporization if the droplets were burning in a stag-
nant atmosphere. In the convective environment of the oil burner,
the burning rate constant k1 can be larger by a factor of about 2 or
3 than the k1 for stagnant conditions and, thus, the radiant energy
flux is probably only about 6 to 9 percent of the total flux required
for vaporization. This is a relatively insignificant percentage; how-
ever, during droplet ignition, just after droplets leave the spray
atomization nozzle, the surrounding gas is cold and the radiant energy
flux must be the major source of thermal energy for ignition of the
droplets. Therefore, the radiant energy fluxes observed are concluded
to have little influence on the combustion processes after ignition,
but to have major influence on the spray droplet ignition processes.
156
-------
Since the radiant energy flux has a strong influence on the spray
droplet ignition process, and since the chamber geometry will, in
turn, influence the radiant energy flux (as well as hot-gas recircu-
lation), oil burner operation in different chambers should be more
consistent if the spark ignition is left on. When the spark ignition
is left on, droplet ignition is less dependent on radiation and hot-
gas recirculation from the rest of the chamber. It should be noted,
however, that consistent operation does not necessarily imply best
operation, and thus there may be more compelling reasons for not leav-
ing the spark ignition turned on, such as reduction of total power
consumption or reduction of nitric oxide emissions.
Carlin Model 250 FFD Burner (6.0 gph)
This burner (shown in Fig. 90) is a somewhat unique design that en-
closes the most upstream portion of the flame in a funnel-shaped
addition on the end of the blast tube. Inside the flame funnel, at
the point of transition from the cylindrical blast tube to the coni-
cal flame funnel, a radially louvered flame retention cone is located.
The air that passes through the swirl-inducing louvers participates
in the combustion that takes place inside the flame funnel, while the
air that passes around the flame-retention device serves to insulate
the large cone from the thermal effects of the combustion which it
contains. The position of the flame-retention cone in the flame fun-
nel is adjustable by means of an escutcheon on the outside of the
burner; this adjustment allows the burner to operate over a fairly
wide oil flowrate range.
The Carlin Model 250 FFD burner was test fired in the 30-inch-diameter,
hot-wall chamber. The maximum recommended firing rate for this burner
is 12 gph; however, with the existing motor/fan combination, it was
not possible to flow sufficient air through this burner to achieve
air-rich combustion, presumably because of the small draft available
from the 30-inch-diameter combustor which has no significant exhaust
157
-------
•4
t
v. •
50
OS
/
rro
•M
-J-i
i
t
t
%.» •'
fL
U
DIMENSIONS-MOOCLS 2SOFFD 5 8SOFFO
(AJI tftmtntiont m tnchit)
45* I 200FF FLANGE
WITH SPACCH
MODEL NO
2V>FFO
C I P,
»«I5 T"'/
02
9%
20 '>i
(a) Overall geometry
(b) Flame funnel detai1
Figure 90. Diagram of the Carlin Company FFD flame funnel
series burner showing the semi-staged combustion
158
-------
stack height. For the 250 FFD, the minimum recommended firing rate
is 5 gph. In this burner, twin oil nozzles are used; however, the
only near-matched nozzles available were a 3-gph, 45 degree hollow-
cone nozzle and a 3-gph, 45 degree solid-cone nozzle. The 250 FFD
burner was fired with these nozzles, which have a total flowrate of
6 gph. The nitric oxide emissions obtained with these nozzles is
shown in Fig. 91. Results are shown for two positions of the flame-
retention cone, the recommended position for 6 gph (1/7-inch forward)
and full forward. The manufacturer recommends positioning the flame-
retention cone at proportionate locations between its full-back posi-
tion (i.e., in the throat of the burner flame funnel) for the lowest,
5-gph flowrate, and full forward (1 inch ahead of full back) for 12
gph. The results shown in Fig. 91 indicate the burner produces less
nitric oxide at the full-forward position than at the recommended
position. At high excess air with the the recommended position, the
flame tended to detach itself and move downstream of the flame cone.
Also shown in Fig. 91 are emission results from the 6-gph Nu-Way, and
the 7- and 12-gph Sun-Ray burners. The 250 FFD burner does not have
as high emissions as the 12-gph Sun Ray, but neither does it have as
low emissions as the Nu-Way burner. Both the Sun-Ray (discussed next)
and the Carlin are flame-retention burners, while the Nu-Way has only
modest peripheral swirler vanes. Since emissions of both the flame-
retention burners are close together at low excess air, while the
Nu-Way is lower, these results support the conclusion that nonflame-
retention burners tend to have lower emissions at low excess air
conditions.
Sun-Ray Model PHC-34 Burner (12 gph)
The Sun-Ray PHC-34 burner is a 5- to 14-gph burner incorporating a
single oil nozzle when used at the nominal oil flowrate of 12-gph
for the reported experiments. This burner has a 4-inch-diameter
conical flame-retention device located inside the 5-inch-ID blast
tube. The spray nozzle is located at the center of the conical
deflector, and there is no choke ring on the blast tube end.
159
-------
2.S
2.0
o
(U
g
w 1.5
u.
O
w 1.0
o
t-l
g
O
0 5
SUN-RAY (12 GPH)
CARLIN 250 FFD (6 GPH)
ESCH. RECOMMENDED POSITION
CARLIN 250 FFD (6 GPH)
ESCH. FULL FORWARD
SUN-RAY .(7 GPH)
NU-WAY (6 GPH)
BACHARACH SMOKE < 1.0
BACHARACH- SMOKE > 1.0
30-IN. DIAMETER HOT-WALL COMBUSTOR
I
1.00 1.10 1.20 1.30 1.40 1.50
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
1.60
Figure 91. Nitric oxide emissions obtained from the Carlin 250 FFD
flame funnel, the Nu-Way Model CO, and the Sun-Ray Model
PHC burners
160
-------
Air velocity vectors were determined using the Sun-Ray burner. These
tests were conducted in an open-air environment. The data are pre-
sented, for the horizontal plane, in Fig. 92 through 94, and the vec-
tors are two-dimensionally displayed in Fig. 95 and 96. At the 1.5-
and 3-inch axial locations, there exist regions in which it was not
possible to measure finite air velocities with the equipment used for
these experiments, and it is concluded that these areas must be slow
moving zones of recirculation caused by the presence of the flame-
retention device in the blast tube exit.
Nitric oxide and smoke emission data for the Sun-Ray burner fired into
the 30-inch-diameter coaxial combustion chamber are shown in Fig. 97
and 98. The burner was tested at two different oil flowrates, using
the same spray nozzle, by adjusting the oil supply pressure. At 7 gph,
the results are very similar to those previously presented for the
6-gph Nu-Way burner while, at 12 gph, the nitric oxide emissions are
approximately 30 percent higher, probably due to the lesser percentage
of heat lost through the chamber walls at the higher flowrate. The
effects of cold- and hot-wall operation are also similar to those same
effects discussed previously in the Nu-Way burner.
Local combustion gas samples were also withdrawn and analyzed for the
Sun-Ray burner operating in the 30-inch-diameter hot-wall chamber.
The stoichiometric ratio, carbon monoxide, and nitric oxide profiles
obtained from these tests are shown in Fig. 99 through 101, as functions
of radial position (r), and axial position (z), relative to the end
of the blast tube centerline. The same data are replotted on a re-
duced scale in Fig. 102 for*comparison with similar data for the same
burner operating under cold-wall conditions. This burner has a flame-
holding cone in the end of the blast tube which serves as a large
zone of ignition for the air/oil spray mixture injected by the burner.
The recirculation pattern caused by this flame cone is surrounded by
an annulus of high velocity-injected air and, thus, should be rela-
tively insensitive (at short z distances) to the temperature of the
161
-------
Pump Side
-3.0
-2.0
-1.0 0
Radius, Inches
Figure 92. Air flow parameters in the horizontal plane at
1.5 inches downstream of the Sun-Ray burner
with no chamber and no oil flow
162
-------
-3-0
-2.0
-1.0 0 1.0
Radius, Inches
2.0
Figure 93. Air flow parameters in the horizontal plane
3 inches downstream of the Sun-Ray burner
with no chamber and no oil flow
163
-------
«, Degrees
-2.0
-1.0 0 1.0
Radius, Inches
2.0
3.0
Figure 94. Air flow parameters in the horizontal plane at
6 inches downstream of the Sun-Ray burner with
no chamber and no oil flow
164
-------
z=6
.OIN.\\\\\
NO OIL FLOW
NO CHAMBER
s=3.0 IN //
1-1.5 IN._J
DOWNSTREAM
MOTOR
SIDE
GEOMETRY SCALE
i
AIR VELOCITY
VECTOR SCALE
lit,
1.0 IN.
60 FT/SEC
IIII, N
PUMP
SIDE
Figure 95. Air velocity vectors in the horizontal plane
for the Sun-Ray PHC-34 burner with no chamber
and no oil flow
165
-------
It
Geometric Scale:
Velocity Scale:
No Chamber
BOTTOM
SIDE
TOP
SIDE
Figure 96. Air velocity vectors in the vertical plane for
the Sun-Ray PHC-34 burner with no chamber and
no oil flow
166
-------
2.0
HOT WALL
11.5 GPH
HOT WALL
7 GPH
COLD WALL
11.5 GPH
COLD WALL
7 GPH
0.5
SUN-RAY MODEL P!!C SUP.NER
30-INCH-DIAMETER CHAMBER
DELAVAN 7.0-45°-B OIL NOZZLE
BACHARACH SMOKE < 1.0
- — — — BACHARACH SMOKE > 1.0
1.0
1.1 1.2 1.3 1.4
STOICHIO>ETRIC RATIO. (AIR/FUEL)/I4.49
1.5
1.6
Figure 97. Comparison of nitric oxide emissions in the flue gas
for the Sun-Rax Model PHC burner operating at 11.5
gph (300 psig oil pressure) and 7 gph (100 psig oil
pressure) in the 30-inch-diameter chamber under
cold- and hot-wall conditions
167
-------
HOT WALL
U)
to
9
8
7
6
5
£
54
2
1
0
1.0
COLD WALL
1.02
1.04
1.06
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
Figure 98. Smoke content of the flue gases for the Sun-Ray Model PHC
burner fired at 11.5 gph in the 30-inch-diameter coaxial
combustion chamber under cold- and hot-wall conditions
168
-------
30 Inch Diameter Hot Wall Combustor
STOICHIOMETRIC RATIO
CARBON MONOXIDE
----- NITRIC OXIDE
z = 6.0 In.
5 10
r, Radial Position, Inches
z » 3.0 In.
200
Q
tu
150
UJ
e
100
so
g
•x.
o
u
—*u
14
H200
o
g
150
u
£
u.
o
100
tu
Q
X
sof
O
0 5 10 14
r, Radial Position, Inches
* = Exhaust Stack Emission Level
Figure 99. Hot-wall, local combustion gas analysis profile for the
Sun-Ray burner at 11.5 gph and a nominal stoichiometric
ratio of 1.03 (single 7-45°-B oil nozzle)
169
-------
5 10
r, Radial Position, Inches
Ol O
Tf Ul
•-H 3
•
-------
O)
•
*fr
e& r*
M O
m
U Q
u
e
CO
30 Inch Diameter Hot Wall Combustor
STOICHIOMETRIC RATIO
CARBON MONOXIDE
NITRIC OXIDE
z = 15.0 In.
150
100
50
*°
14
Figure 101.
5 10
r, Radial Position, Inches
= Exhaust Stack Emission Level
Hot-wall, local combustion gas analysis profiles for the
Sun-Ray burner at 11.5 gph and a nominal stoichiometric
ratio of 1.03 (single 7.0-45°-B oil nozzle)
u.
o
A
i
z
-------
| Overall Stoichiometric Ratio • 1.03 — 30-Inch Diameter Chamber
5 i
SR
0
(
5
SR .
0
200-
1 z-3
™-_jir —
0
) ' r ' IS C
200.
z o 6
0
Cold Wall
Hot Wall
z = 3
) ' r ' 15
/M z = 6
5 -.
NO '
5.
0
/
D
{
z = 3
' r ' 1
s
15
5.
SR.
0
z = 9
IS
200
coj
0
Z « 9
IS
5.
NO'
0
z » 9
15
5
SR,
0
z - 12
r ' 15
200
CO.
0
* - i2
0 r 15
z = 12
5
SR.
0
z - IS
IS
z <= 15
NO
z - 15
15
Figure 102. Comparison plots of local combustion zone sampled
Stoichiometric ratio (SR, air/14.49 fuel), carbon
monoxide (gm CO/kg of fuel burned), nitric oxide
(gm NO/kg of fuel burned), as a function of radial
location (r = inches) and axial location (z = inches)
(Sun-Ray burner, 115 gph)
172
-------
recirculated gases and, hence, insensitive to whether hot- or cold-
wall conditions are being tested. The gases recirculating behind the
flame cone are a main source of ignition for the injected air/fuel
spray mixture and, in agreement with the fact that these flame cone
recirculating gases should be insensitive to wall temperature, the
stoichiometric ratio profiles for hot- and cold-wall conditions show
a high degree of correspondence to one another, particularly near the
centerline, r = 0. The carbon monoxide profiles show few differences,
except at z = 3 inches, where the outer radial limit of the active
combustion zone (i.e., high CO) is higher in the hot-wall case. The
nitric oxide profiles are generally higher for the hot-wall case, as
compared to the cold wall, except for a portion of the profile at
z = 15 inches. With this burner, as with the Nu-Way burner, the nitric
oxide content of the recirculating gases, out along the chamber wall,
is about 20 percent lower than the content of the flue gases. Here
again, the difference is probably due to the lack of recirculation of
the nitric oxide-rich gases in the core of the flow.
INTERPRETATION OF COMMERCIAL BURNER RESULTS
In this section, limited portions of the experimental results presented
previously are discussed. The objective of this discussion is to make
the point that burners tend to operate in one or the other or a com-
bination of two fundamental modes, which are referred to herein as the
plug-flow reactor and the well-stirred reactor modes. The experimental
measurements of local flame-zone pollutant concentrations are inter-
preted to show that burners that tend to behave as plug-flow reactors
produce lower nitric oxide emission levels when less than 80-percent
excess air is utilized, while those that tend toward the well-stirred
reactor behavior produce lower nitric oxide emission levels only with
the impractically high excess air levels greater than 80 percent.
Finally, because of these interpretations, it is concluded that burner
designs should be of the plug-flow type to achieve low nitric oxide
emissions. This requires that burners be of the nonflame-retention
173
-------
type and well designed so that they can achieve: (1) uniform dis-
persion of the fuel spray in the air flow, and (2) eliminate or mini-
mize turbulent mixing in the combustion zones. These conclusions are
reached by reference to previous results, as described in the follow-
ing paragraphs.
Velocity vector and composition profile maps are shown in Fig. 67
through 73 for a Mite burner. The combustion gas velocity vectors
measured at three axial locations downstream of the burner's blast
tube discharge are shown in Fig. 69 and 70. Also shown in Fig. 69
and 70 is a diagram of the blast tube end and the combustion chamber.
The large disturbances in the flow pattern that result from the pres-
ence of the multiple-blade swirler device in the end of the blast
tube are shown later to have an effect on nitric oxide formation.
Figures 71 through 73 show combustion gas composition profiles measured
at the same three axial locations at which the Mite velocity profiles
were obtained. The most interesting features in Fig. 71 through 73
is the presence of the prominent peaks of nitric oxide content that
happen to be located at the flow disruptions caused by the vaned-
swirler device. The local stoichiometric ratio near the nitric oxide
peaks is near unity. The results in Fig. 71 through 73 suggest that
thorough mixing (due to the vaned swirlers) in regions of near stoi-
chiometric combustion leads to excessive nitric oxide formation.
Velocity vector and composition profile maps were shown in Fig. 46
through 56 for the 55-J burner. The construction of the 55-J burner
blast tube, as shown in Fig. 53, is considerably different from the
flame-retention-type burner. There are swirler vanes, but these are
located inside the blast tube choke ring at diameters larger than the
inside diameter of the choke ring, so that they do not lead to major
disruptions of the flow pattern. Note that the flow pattern is
slightly nonsymmetrical, with a larger velocity gradient on the motor
side of the velocity profile than on the pump side. This nonsymmetry
is caused by the blower arrangement and various components inside the
blast tube upstream of the components illustrated in Fig. 53. Note
174
-------
also, in Fig. 54 and 55, that there is a definite tendency for the
local nitric oxide content of the combustion gases to be higher on
the side near the largest velocity gradients. This, once again, infers
that strong mixing (a result of the large velocity gradients) leads to
excessive nitric oxide formation. In this case also, the nitric oxide
formation is highest in the strong mixing, nearly stoichiometric regions.
Similar results were also obtained with the larger burners. Figures
95 through 101 illustrate results obtained with the 12-gph Sun-Ray
burner, for example. Because of the high flowrates involved, it was
more convenient to measure velocity vectors only under a no-combustion
condition. The velocity vector profiles shown in Fig. 95 and 96 in-
clude large regions near the blast tube with no vectors shown. In
those regions, the velocities were either too small for accurate
measurements, or directed backwards (recirculation) so that they were
not measurable with the apparatus used. Therefore, there appears
from Fig. 95 and 96 to be a large region of low velocities and recir-
culation caused by the flame-holding cone located in the end of the
blast tube. The combustion gas composition profiles shown in Fig. 99
through 101 indicate very high nitric oxide content in that recircula-
tion zone. Once again, these results suggest that regions of strong
mixing are not desirable for minimization of nitric oxide formation.
The velocity vector and gas composition results presented in this sec-
tion all suggest that vigorous mixing at near stoichiometric conditions
should be avoided if nitric oxide formation is to be reduced. Carried
to an extreme, it might logically be concluded that a burner that sim-
ulates a plug-flow reactor might be expected to have lower nitric oxide
emissions than one that simulates a well-stirred reactor.
A plug-flow reactor is characterized by the manner in which the react-
ants are allowed to react. In a plug-flow reactor, any particular
differential volume of reactants is allowed to carry out the reaction
in a timewise progressive manner without the addition of any other
175
-------
components (i.e., no fresh reactants and no products from outside the
differential volume) throughout the course of the reaction. In a two-
phase, liquid/gas reaction, such as the vaporization and combustion of
oil spray in air, strictly speaking, the oil droplets originally asso-
ciated with a given differential volume of air should remain with that
differential volume of air until the droplets are completely vaporized
and the combustion process is completed. However, as far as the reac-
tion chemistry is concerned, it is of no significant consequence if
those specific droplets are removed from the differential gas volume
and replaced by other droplets. In other words, an oil spray combus-
tion process can be considered to be essentially a plug-flow-type
process, in spite of gross changes in the spray location, as long as
no intermixing of the gas phase occurs during the combustion process.
That is, the differential reaction volume starts out as a pocket of
pure air, to which modest amounts of fuel vapor and heat are added by
droplet vaporization and heat conduction. This leads to ignition, and
then the process proceeds with fuel vapor addition to the air pocket,
by vaporization, and reaction of that fuel vapor with the available
oxygen in the air pocket. The process continues until no more fuel
is available to be added. If the overall process is near stoichio-
metric (S.R. = 1.0), the end point for each differential volume of
•gas should be such that little or no unreacted oxygen will remain.
A well-stirred reactor is also characterized by the manner in which
the reactants are allowed to react. In a well-stirred reactor, reac-
tants are added quickly and uniformly to a relatively large volume of
material in which the reactions are already nearly complete. In an
oil burner operating in the well-stirred mode, air and oil spray are
added to a strongly stirred volume of material consisting primarily
of reaction products. If the overall process happens to be operating
air rich, this volume of products will contain excess oxygen and very
little fuel vapor whereas, if the overall process happens to be oper-
ating fuel rich, the volume of products will contain significant
amounts of fuel vapor or fuel pyrolysis products and very little oxygen.
176
-------
The primary chemical difference between a well-stirred reactor and a
plug-flow reactor is, therefore, the bulk gas composition under which
the reaction takes place. In a well-stirred reactor, the chemical
reactions take place entirely in bulk gas having a composition equiva-
lent to the combustor exhaust whereas, in the plug-flow reactor, the
chemical reactions take place throughout the entire continuous spectrum
of bulk gas composition, ranging from essentially pure air to the com-
bustor exhaust composition. For chemical species that reach equilib-
rium quickly, such as H-0, and even CO/CO , the concentrations in the
exhaust will be very near the equilibrium amounts no matter whether
the reaction is carried out as a plug-flow or a well-stirred process.
However, the mode of reaction, be it plug flow or well stirred, would
be expected to have a strong influence on the amount of particular
reaction products in the exhaust that have slow kinetics and are char-
acteristically found in concentrations far from equilibrium (e.g.,
NO). This expectation constitutes a primary thesis of this section,
i.e., the NO produced by an adiabatic combustion chamber oil burner
is largely dependent on whether the burner tends to operate as a plug-
flow burner, or as a well-stirred burner, or as some intermediate
combination of the two modes.
Nitric oxide is generally thought to be formed by either of two simi-
lar processes which are labeled "prompt" and "thermal." Thermal NO
is produced by allowing combustion products to exist for a period of
time without significant heat loss, so that reactions such as the
following can occur:
N2 + 0 ^ NO + N
0 + N ^ NO + 0
N + OH ^ NO + H
where, for the thermal NO process, the species such as 0 and OH exist
in chemical equilibrium with the other combustion products.
177
-------
Prompt NO, on the other hand, is formed by chemical reactions similar
to those above, but under conditions in which the active species such
as atomic oxygen (0) exist in super-equilibrium concentrations. These
super-equilibrium concentrations exist because the species are reaction
intermediates. That is, the very active intermediate species such as 0
exist for short durations in super-equilibrium amounts because many
chemical bonds are being broken and formed in the process of oxidizing
the available fuel. The rate at which these super-equilibrium amounts
of active species are destroyed by non-objectionable reactions (i.e.,
those which do not produce NO) must vary significantly depending on
the composition and temperature of the bulk gas in which they are found.
The amount of prompt NO formed per unit of fuel oxidized should not
depend on the total bulk gas product residence time in the combustion
chamber (as it does for thermal NO) because the super-equilibrium
amounts of active intermediate species do not exist for time periods
anywhere near comparable to chamber residence times.
A separate mechanism for the formation of NO, with chemical kinetics
which are far different from the kinetics in the equilibrium bulk gas,
can occur in the boundary layer flame fronts surrounding individual
droplets, or in the flaming wakes of individual droplets. In these
regions, localized temperatures are near the stoichiometric, adiabatic
flame temperature, depending somewhat on the surrounding bulk gas tem-
peratures. At these high, localized temperatures, NO kinetics are
more rapid and large amounts of NO might be formed. This mechanism
might be referred to as a "quick" mechanism, since the chemical species
exposed to the high temperature are exposed only briefly, as they pass
through the localized flame fronts. The amount of NO formed by this
mechanism is dependent on the fraction of the combustion volume which
is at the high temperature which, in turn, is proportional to the
amount of fuel being burned. Thus, NO being formed by this quick mech-
anism is dependent on the amount of fuel which has been burned, and
not at all on the bulk gas residence time after completion of burning.
178
-------
Formation of thermal NO, in contrast to the quick and prompt NO's dis-
cussed above, is primarily dependent on temperature and residence time
(i.e., time for reaction) after the bulk gases have reached their maxi-
mum temperatures. Thus, the amount of prompt NO and quick NO will both
be functions of the amount of fuel burned and somewhat dependent on
bulk gas composition and temperature (as it affects the super equilib-
rium oxygen atom lifetimes, and the local flamefront temperature), but
not dependent on bulk gas residence times. The amount of thermal NO,
on the other hand, is dependent on bulk gas residence times and not
particularly on the amount of fuel being burned.
Assuming for the moment that the NO with which we are dealing is only
quick and/or prompt NO, then it must be possible to relate NO forma-
tion in plug-flow combustion processes to NO formation in well-stirred
combustion processes. As stated above, the amount of NO formed per
unit mass of fuel oxidized should depend only on the bulk gas temper-
ature and composition at which the oxidation takes place. But, for
an adiabatic process at near-equilibrium conditions for the major
components, these are continuous functions of the bulk gas stoichio-
metric ratio. Therefore,
(ID
where
W = amount of prompt NO formed
WC1ICI = amount of fuel oxidized
rllcL
SR = stoichiometric ratio of the bulk gases
f(SR) = a continuous function of stoichiometric ratio, to be
defined experimentally
Now, in a well-stirred reactor, the stoichiometric ratio of the bulk
gases is constant throughout the reactor at the value of the exhaust
179
-------
composition (SR^g) where the subscript WS refers to well stirred. The
function f(SR) can thus be moved outside the integral sign and integra-
tion of Eq. 11 under these conditions yields:
r(VWw.S (WFUEL5WS
/dWNQ = / f(SR1B)dWpUBL = f(SRws) J dWFU£L (12)
In a plug-flow reactor, the stoichiometric ratio of the bulk gases
varies as the fuel is oxidized, so that SR = WATi/(14'49 WFUEL^' so
that:
dWNO ' fdWFUEL
or
f(WAIR/14'49 W
0
and, in this case, the function cannot be moved outside the integral
sign.
Now, according to the discussion presented just prior to Eq. 11, the
function f(SR) used in Eq. 12 through 15 remains the same regardless
of whether the reactor is a plug-flow or a well-stirred type. In this
case, the function can be evaluated from well-stirred reactor NO emis-
sion data by algebraic manipulation of Eq. 13 to yield:
f(SR)
(16)
SR
180
-------
Or, it can be evaluated from plug-flow reactor data by differentiation
of Eq. 15 to yield:
(17)
- CD
LUG
f(SR) = "° PLUG
FUEL^PLUG „„ _ SR
Having used Eq. 16 to obtain the required function from well-stirred
combustor NO emission data, it should be possible to apply that func-
tion to Eq. 15 to predict the NO emission data for a plug-flow combus-
tor. Conversely, having obtained the function from differentiation
of plug-flow reactor data according to Eq. 17, it should be possible
to predict well-stirred combustor NO emissions. As shown below, this
has been successfully done during this program with emission data from
burners operating predominantly in the plug-flow mode and the others
operating in the well-stirred combustor mode. The important factor
here is that the function f(SR) has been experimentally determined,
so that it is possible to define: (1) the conditions for which plug-
flow combustors are superior to well-stirred combustors, and vice-
versa; and (2) the lowest level of NO emission that it is possible to
obtain in adiabatic combustors. This informativ,. allows one to decide
which type of combustor development to pursue, and it defines the NO
level at which further refinements in burner design for adiabatic com-
bustors will not result in further significant decrease in NO emission.
It should be emphasized that the above arguments are based on the use
of an adiabatic combustion system, either plug flow or well stirred,
where there is no significant heat removal during the course of com-
bustion, heat loss by radiation or convection to combustion chamber
walls, or effective heat loss by techniques such as flue gas recircu-
lation. As such, the above arguments are applicable to luminous-flame
burners fired into refractory combustion chambers which are a principal
subject of this research program. On the other hand, the above argu-
ments are not applicable to water- or air-cooled combustion chambers
for which significant flame-zone heat losses must be considered.
181
-------
One condition for the applicability of the arguments related to Eq. 11
through 17 was that "prompt" or "quick" rather than "thermal" NO for-
mation mechanisms are predominant. The reactions for production of
thermal NO depend largely on the presence of atomic oxygen, which exists
in adiabatic equilibrium combustion products in significant amounts
only when less than 35 to 40 percent excess air is utilized. With 40
percent or more excess air, the thermal NO mechanism would be expected
not to produce significant amounts of NO from equilibrium combustion
products due to the absence of 0, leaving only the prompt mechanism
for NO formation. Therefore, the arguments with respect to NO forma-
tion, as related to Eq. 11 through 17, should be valid for greater than
35 to 40 percent excess air, or under conditions of minimum chamber
volume (i.e., low residence time at peak temperature), or with slight
radiant heat loss (slightly reduced flame temperature and hence lower
0 concentrations). On this basis, the relations of Eq. 11 through 17
might be expected to be valid to as low as, say, 10- to 15-percent
excess air under conditions of low chamber residence time. The exper-
iment interrelationship of well-stirred and plug-flow combustor NO
emission data, as described below, has been attempted only in the 35-
to 100-percent excess air range.
If an oil burner were to act as a true plug-flow reactor, the process
could be visualized as a plug of air to which fuel vapor is continu-
ously added (by vaporization from droplet) and reacted. As the process
continues, the buildup of combustion product concentrations would be
expected to increase monotonically until a near-stoichiometric condi-
tion is reached, at which time the process would be considered com-
plete. For a practical oil burner, the entire combustion might not
be a single plug-flow combustion process; however, if the air flow
were separated into several different stream tubes, each of the stream
tubes might-be considered to be individually undergoing a plug-flow
combustion process if turbulent mixing were not excessive. Then, if
local combustion gas composition profiles were measured, each combus-
tion gas sample could be considered to be a sample from a typical plug-
flow combustor at a particular degree of completion.
182
-------
Among the commercial burners tested, the 55-J has the least vigorous
mixing, and the resulting combustion would be expected to provide the
best simulation of a plug-flow reactor. Even though the flow is in
the turbulent regime, the results obtained from combustion profile
sampling of the 55-J might reasonably be expected to exhibit streatube
flow characteristics.
Combustion zone sampling data for the 55-J, including the data from
Fig. 54 are shown plotted in Fig. 103 in such a manner that consistent
trends should be observed if the streamtubes from which the' samples
were obtained actually simulate plug-flow reactor behavior. Figure
103 shows the weight of nitric oxide per unit weight of air as a func-
tion of the mass of fuel per unit mass of air. The data in Fig. 103
include combustion sample analyses from throughtout the combustion
zone of the 55-J. Since the sample analysis train does not permit
liquid fuel or fuel vapor to enter the sample analyzers (both are
removed by passage of the gas sample through a cold trap), the weight
of fuel per unit weight of air for Fig. 103 has been calculated from
the carbon dioxide and carbon monoxide concentrations and, hence, the
fuel referred to on the abscissa of Fig. 103 is burned fuel. The plug-
flow combustion process starts with pure air, followed by monotonic
increase in fuel/air ratio, which can be visualized as moving right-
ward from the origin on the abscissa of Fig. 103.
The data of Fig. 103 show a relatively consistent trend. If the data
truly describe the characteristics of a plug-flow combustion process,
then it is apparent that little or no nitric oxide is generated when
the plug composition passes from an abscissa value of 0.01 to 0.04
gm fuel/gm air. Most of the nitric oxide is formed at 0.05 to 0.07
gm fuel/gm air, with some apparently formed'from 0.0 to 0.01. The
curve drawn in Fig.103 is related to the function f(SR) from Eq. 11
through 17, as shown by the equation in the figure. The curve drawn
in Fig.103 is purposely placed near the lower limit of the experimental
183
-------
250
oo
200
USD
< O
£- ISO
X
C3
.
z o
•^ 3e2 100
so
o
o
DATA OBTAINED BY LOCAL SAMPLING FROM COMBUSTION
TESTS IN WHICH THE OVERALL BURNER STOICHIOMETRIC
RATIO WAS 1.54 AND 1.29, WITH SAMPLES TAKEN AT
1.50, 3.00, and 6.00 INCHES DOWNSTREAM OF THE
BLAST TUBE, AT RADII 0.0 TO 3.5 INCHES FROM THE
IMAGINARY EXTENSION OF THE BLAST TUBE AXIS.
'WFUEL/WAIR
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
LOCAL BURNED FUEL TO TOTAL AIR RATIO, WFUE|_/WA|R
Figure 103. Nitric oxide content of locally sampled combustion gases,
as a function of combusted fuel content, for the ABC 55-J
burner (which has relatively streamlined flow)
O.OB
-------
data because of the predominantly positive value of the second deriva-
tive of the curve. (That is, if two equal weight strearatubes having
fuel/air ratios of 0.03 and 0.07 were to become mixed by turbulent
actions, the mixed gas fuel/air ratio would be 0.05; but, the weight
of nitric oxide per unit weight of air in the sample would be about
65 x 10 rather than the value of 40 x 10" which would result if
a single plug-flow streamtube reached the fuel air ratio of 0.05 by
the plug flow combustion process. Since the turbulent mixing which
does occur tends, by this effect, to elevate the data points, the
true curve for plug-flow combustion must lie near the lower limit of
the data, as it is drawn in Fig. 103.)
The large scatter in Fig. 103 is not surprising, since the flow pro-
duced by the burner is not truly plug flow, but somewhat turbulent.
The other burners tested, which have devices tending to promote tur-
bulence, produced data trends similar to those of Fig. 103, but with
relatively larger scatter. If it were important to obtain good data
representative of plug flow, a burner operating in the laminar flow
regime could be designed, and the gases sampled from its combustion
zone would probably have much reduced scatter when plotted according
to the format of Fig. 103.
In a well-stirred process, the gas composition in the combustion cham-
ber is stationary as fuel is added, and the amount of nitric oxide
formed per unit of fuel added should (according to Eq. 16 and 17)
be proportional to the slope of the curve representing the data in
Fig. 103. The large slope in Fig. 103 at the stoichiometric condition,
therefore, indicates that a well-stirred reactor operating at near
stoichiometric would produce large amounts of nitric oxide. This is
consistent with the trends discussed in relation to the combustion
zone flow pattern and gas sample concentration data presented earlier
in which high nitric oxide content was found at near-stoichiometric,
"well-stirred" zones. It is noticeable that the slope of Fig. 103 is
185
-------
near zero at the intermediate range of the abscissa. Obviously, a
well-stirred reactor operating in this more air-rich zone would produce
relatively low nitric oxide emissions.
A special oil burner, discussed in the nonconventional burner section,
was fabricated to test the characteristics of the well-stirred burner
concept. This burner achieves the well-stirred characteristics by vig-
orously swirling the air before it is injected into the combustion cham-
ber. The air swirl is induced by rotating a six-vane paddle wheel in
the end of the burner at 3450 rpm. The flue gas emissions data obtained
with this intense swirl, "well-stirred" burner are shown in Fig. 104.
The curve was obtained by differentiating the corresponding portion of
the plug-flow curve shown in Fig. 103, and multiplying the local burned
gas fuel/air ratio, indicating that the plug-flow and well-stirred
reactor data are related to each other by the function f(SR) as they
should be if the above-described assumptions are true. The function
f(SR) has been evaluated numerically and, in terms of fuel to air
weight ratio, it is:
dW
dw N° = f(SR) = 26.126 - 34.951 (SR) + 16.152 (SR2) - (18)
FUEL
2.5369 (SR3) (gm NO/kg fuel); 1.0 < SR < 4.6
If the curve from Fig.104 were superimposed on Fig. 103, the crossover
between the plug-flow and well-stirred reactor curves would occur at
about 0.04 gm fuel/gm air (i.e., about 80 percent excess air). To the
left of the crossover point, the well-stirred reactor design produces
lower nitric oxide emissions while, to the right of the crossover point,
the plug-flow reactor has lower nitric oxide emissions.
The crossover point between the two reactor types occurs at a rela-
tively high excess air level. At high excess air levels, the flame
temperature is reduced, making it more difficult to transfer heat from
the combustion gases to the heated media and, therefore, generally
186
-------
200r-
150
NO
o
ec
<
100
x
50
0.02
• EXPERIMENTAL DATA
^DETERMINED FROM CURVE IN FIG. 103
§
O
o
CO
0.03
0.04
0.05
0.06
0.07
Wc,,ri/W.,D = EXHAUST GAS BURNED FUEL/AIR RATIO
rUtL nlK
Figure 104. Nitric oxide content of flue gases produced by
the well-stirred burner
187
-------
lowering the overall furnace thermal efficiency (typically by 10 to
15 percent at 80-percent excess air versus 10-percent excess air).
This loss in efficiency is not tolerable in existing furnaces and,
hence, low excess air levels must be used; therefore, burner designs
which promote plug-flow mode behavior must be selected. With this
selection, at 10-percent excess air (WMCIWATD = 0.063), Fig. 103
indicates that about 70 x 10 " grains of nitric oxide per gram of air,
or about 1 gm NO/kg fuel will be produced. For adiabatic combustors
(i.e., uncooled combustion chamber systems with little radiant or
convective heat loss, and no cooled flue gas recirculation) , this
1 gm NO/kg fuel level is about the lowest that should be expected
since, to achieve it, the oil burner must operate as a plug-flow
burner, uniformly for all of the flow it passes. In practice, non-
uniformities and some turbulent mixing will be difficult to avoid,
and the ideal goal of 1 gm NO/kg fuel for an adiabatic combustor might
be expected to be somewhat exclusive.
Considerations of furnace efficiency, as discussed above, dictate that
all burners operate at low excess air. As also discussed above, at
low excess air, burners operating in the plug-flow mode have lower
nitric oxide emissions than burners operating in the well-stirred mode.
Therefore, the burner optimization experiments described in the follow-
ing section were conducted with burners that tend to operate in the
plug-flow mode. This eliminated the possibility of using flame-
retention devices or other devices that tend to promote vigorous tur-
bulence in the air injected through the blast tube.
BURNER GEOMETRY OPTIMIZATION STUDIES
The results discussed in the previous section strongly suggest that
plug flow-type burners have the greatest likelihood of operating with
low nitric oxide emissions in refractory- lined combustion chambers.
Furthermore, it is apparent that burners that do not utilize flame-
retention devices are most likely to produce plug-like flow. The
188
-------
burner geometry optimization studies described in this section have,
therefore, made use of versatile geometry burners that do not incorpo-
rate flame-retention devices. Instead, they have only peripheral
swirlers which resemble those found on the 55-J-l and model CO burners
discussed previously. The variable angle swirlers allow a controlled
amount of turbulent mixing to be induced into the flow, varying from
zero to a moderate amount.
Two versatile geometry burners were constructed for this study, one
for application from 0.75 to 3 gph, and one for application from 3 to
12 gph. The 3- to 12-gph version is shown in Fig. 105. Geometry vari-
ation effects on pollutant emissions were studied at 0.75 and 3 gph
with the smaller version, and at 7 and 12 gph with the larger version
of the versatile burner. The most detailed studies were conducted at
the 12-gph flowrate.
A photograph of the large, 3- to 12-gph versatile burner is shown in
Fig. 105. It has a six-leaf, iris-type variable choke plate which
results in a hexagonal-shaped outlet that can be varied in area from
about 3 to 15 sq in. while the burner is running. The choke plate
settings, as reported herein, are given in terms of the equivalent area
circle diameter. The versatile burner also has six variable angle
swirler blades, which can be seen just inside the blast tube exit in
Fig. 105. These swirler vanes are mounted so that their angle relative
to the blast tube axis can be varied from 0 to 60 degrees while the
burner is running. In the earliest studies with the 1- to 3-gph versa-
tile burner, the variable angle swirler vanes were relatively short in
the radial direction, with the innermost edges located at a radius of
1.5 inches relative to the blast tube centerline. Experimental results
showed these short swirler vanes to have no significant effect on emis-
sions. For all subsequent experiments, the swirler vanes were made
taller so that their innermost edges were located at a radius of 0.6
189
-------
I—
o
o
5AD26-7/17/73-S1B
Figure 105. Overall photograph of the 3- to 12-gph versatile research burner
-------
inch from the blast tube axis (this radius is just slightly larger than
the radius of the oil spray nozzle). The experimental results with
these tall swirler vanes showed significant effects of swirl angle on
emissions.
Photographs of the head ends of the 3- to 12-gph and 1- to 3-gph versa-
tile burners are shown in Fig. 106.
The 12-gph studies were conducted using the 30-inch-diameter coaxial
combustion chamber 35 inches in length. The chamber was internally
lined with a 1/2-inch-thick layer of Pyroflex insulation. The exhaust
from the coaxial combustor was passed through the mixing and exhaust
section to thoroughly mix the gases to provide a sample having average
composition for the analyses. The oil spray nozzles used for these
studies were all Delavan brand, either A = hollow cone spray, or B =
solid cone spray, with spray angles of 45, 60, or 90 degrees. The oil
nozzles are nominally rated for flow at 12 gph; however, a nominal 7-
gph nozzle, when operated at an elevated pressure of 300 psig, also
flows at about 12 gph.
Experiments conducted with the versatile burners in coaxial chambers
generally had very low emissions of unbumed hydrocarbons and carbon
monoxide when operated at those conditions that yielded low smoke;
however, low nitric oxide emissions did not necessarily coincide with
the low smoke emissions. It should be noted that, in some instances,
significant smoke was observed in the absence of significant hydrocarbon
or carbon monoxide emissions. This may have been due, for example, to
impingement of fuel on the chamber walls followed by formation of smoke,
vaporized hydrocarbon, and carbon monoxide; subsequently, the hydro-
carbons and carbon monoxide may have become oxidized, but the smoke
(which is more difficult to oxidize) persisted into the exhaust. How-
ever, the reverse was not generally observed, i.e., significant hydro-
carbons or carbon monoxide in the absence of smoke. Therefore, the
191
-------
1- TO 3-GPH HEAD
(COMPLETE)
6- TO 12-GPH HEAD
(LESS VARIABLE
SWIRLERS)
6- TO 12-GPH VARIABLE CHOKE,
FULL OPEN
6- TO 12-GPH VARIABLE CHOKE,
FULL CLOSED
Inches
Inches
Figure 106. Photographs of the variable geometry burner heads of both
the 1- to 3-gph and the 6- to 12-gph versatile burners
192
-------
problem of selecting the optimum burner geometry reduces to selecting
those conditions where low smoke emissions coincide with low nitric
oxide emissions.
The results of the 12-gph studies are presented in Fig. 107 through 110.
Figure 107 shows the results of an attempt to determine the effects of
dropsize by using two nozzles of the same spray pattern, but requiring
different operating pressures to obtain the same flowrate. Figure 107
also shows the effects of choke diameter and air swirl angle. Figure
108 is similar to Fig. 107, except it is for wider angle spray nozzles.
Figure 109 is also similar to Fig. 107, except for spray angle, and it
also includes the effects of the hollow cone versus the solid cone spray
nozzle. Figure 110 shows the effects of oil spray nozzle recess with
respect to the blast tube exit. Figures 107 through 109 indicate that
dropsize has little or no significant effect on the minimum achievable
NO levels, but that the smaller dropsizes (i.e., those obtained from
the highest operating pressure nozzles) have a lesser tendency to pro-
duce smoke.
According to Fig. 107 through 109, increasing the air swirl vane angle
results in an almost monotonic increase in the NO emissions under the
no-smoke condition. This result is not surprising, since the commer-
cial burner studies described in the previous section indicated that
well-stirred combustion at low excess air levels tends to produce
higher levels of NO than plug-flow combustion. Increasing the air
swirl angle obviously makes the combustion more like the well-stirred
process and, therefore, leads to the monotonic increase in NO. At very
low swirl angles, there is a definite tendency toward the formation of
smoke. Apparently, some degree of stirring (by means of the air swirl
vanes) is necessary to achieve sufficiently uniform flow to avoid fuel-
rich zones of combustion where smoke is produced. Therefore, the opti-
mum air swirl vane angle obviously becomes a compromise between an angle
sufficiently large to eliminate smoke, but as low as possible to minimize
193
-------
7-4S°-B NOZZLE
300 PSIG. 12 GPH
gm NO
kgm FUEL
ID
20 40
SKIRL ANGLE, DEGREES
VERSATILE BURNER RESULTS
30 INCH HOT WALL CHAMBER
CHOKE
.DIA.
,3.50"
1.00"
1-^4.50"
60
12-45°-B NOZZLE
100 PSIG, 12 GPH
NO
gm FUEL
2
1
0
(
^ ^ ^
X
^ ^ '
~=^=
-
^-
H
CH
" DI
^4
e 4
OKE
A.
.00"
.50"
) 20 40 60
SWIRL ANGLE, DEGREES
---- SMOKE > 1 . 0
SMOKE < 1.0
Figure 107. Versatile burner flue gas emissions obtained with 10-percent
excess air in the 30-inch-diameter hot-wall chamber
(45° nozzles)
-------
7-90 -B NOZZLE
300 PSIG, 12 GPH
ID
cn
20 40
SWIRL ANGLE, DEGREES
VERSATILE BURNER RESULTS
30-INCH HOT-WALL CHAMBER
60
12-90 -B NOZZLE
100 PSIG, 12 GPH
01
u,
u.
o
1
ta
o
t—i
x
o
a:
20 40
SWIRL ANGLE, DEGREES
BACHARACH SMOKE < 1
— — — — — BACHARACH SMOKE > 1
60
Figure 108. Versatile burner flue gas emission obtained with 10-percent
excess air in the 30-inch-diameter hot-wall chamber
(90° nozzles)
-------
to
(VERSATILE BURNER RESULTS
30 INCH HOT WALL CHAMBER
7-60°-B NOZZLE 12-60°-B NOZZLE
300 PSIG, 12 GPH 100 PSIG. 12 GPH
3
2
1
NO
e
0
-'--
-— -•
~*^
- •"^"^"^^"^
x1
3
CHOICE
DIA.
—4.00
IX
Z
'., 3. 75
SS 4. SO
NO
fr
0
s-^
.j^Z
^~~^
7-60° -A NOZZLE (A-Hollow Cone)
300 PSIG. 12 GPH
3
CHOKE z
. — 4.50
*^
~^J.75
i
0 20 40 60 0 20 40
SKIRL ANGLE, DEGREES SWIRL ANGLE. DEGREES
s
MOKE > 1.0
MOKE <_ 1.0
NO
£B
kg
0
___
— ^ •"
. — — '
^-^
x,
CHOICE
PTA.
x^~
4.0O
60 0 20 40 60
SWIRL ANGLE, DECREES
Figure 109. Versatile burner flue gas emissions obtained with 10-percent
excess air in the 30-inch-diameter hot-wall chamber
(60° nozzles)
-------
VERSATILE BURNER
4.0-Inch-Diameter Choke, 25° Swirl Settings
7-60 -B Nozzle, 300 Psig. 12 GPH Flowrate
30-Inch-Diaraeter Hot-Wall Chamber
to
0
ULJ
g 2
i
u.
s
§ 1
0
•
MBM^MM
V.
,^— -
-1.0 -0.5 0.0 +0.5 +1.0
NOZZLE RECESS, INCHES
u.
o
§1
i
SUN-RAY 7-45 -B. 12 GPH
I
VERSATILE
1.0 1.1 1.2
STOICHIOMETRIC RATIO,
1.3
(AIR/FUEL)/14.49
1.4
——^— BACHARACH SMOKE < 1
.... BACHARACH SMOKE > 1
Figure 110. Versatile burner flue gas emissions obtained with 10-percent
excess air in the 30-inch-diameter hot-wall chamber
-------
NO formation. Examination of Fig. 107 through 109 suggests that an air
swirl vane angle of about 25 degrees is a reasonable compromise between
smoke and nitric oxide.
The effects of choke diameter at 12 gph, as shown in Fig. 107 through
110, are not totally unambiguous. However, the 3.50-inch choke diameter
(Fig. 107) produced relatively high NO emissions, and the 4.5-inch choke
diameter also produced somewhat high NO emissions, and/or required a
large air swirl angle to eliminate smoke. The 3.75- and 4-inch choke
diameters appear, therefore, to be the best compromise.
The effects of spray nozzle recess, as shown in Fig. 107, were not sig-
nificant, and the effects of stoichiometric ratio, also shown in Fig.
110,were minimal with respect to nitric oxide. Also shown in Fig. 110
are NO emissions from a commercial 12-gph burner tested under similar
conditions, indicating the optimum adjustment of the versatile burner
leads to the lower NO emissions.
In Fig. 107 through 110, it is interesting to note that the nominal level
of 1 gm NO/kg fuel appears to be a lower barrier, as suggested by the
analyses described in the previous section. This lends credence to
the conclusion that 1 gm NO/kg fuel is the nominal minimum that can
be achieved by adiabatic, near-stoichiometric oil spray/air combustion
without smoke. A lower value should, of course, be possible when air-
cooled or hydronic combustion chambers are utilized, since combustion
gases cooled by contact with the chamber wall can be aspirated into
the air flow and, also, the flame temperature is lowered due to radia-
j>
tion losses from the flame to the cool wall. These cooling methods,
however, are not possible with refractory wall combustors.
Using the same 30-inch-diameter coaxial combustor, versatile burner
experiments were conducted at 7 gph. The results of these experiments
are shown in Fig. 111. Choke diameters of 3 to 3.5 inches and an air
198
-------
ID
IO
3
2
NO
kgn
1
0
C
7-4S°-B NOZZLE
100 PSIG, 7 GPH
—
— i
_,
-=-±:
\Z
^
^
^*_
s^~*^
CHOKE
DIA.
3.5
*•—- j .1
v
N4 «
3
2
NO
*m-
kgm
1
0
VERSATILE BURNER RESULTS
30-INCH DIAMETER HOT WALL CHAMBER
7-60°- B NOZZLE
100 PSIG, 7 GPH
rJ:--
) 20 40 60 0
SWIRL ANGLE, DECREES
=.--=•'•
^~
S^
^
_ *^.'
CHOICE
DIA.
^ t
,4.0
,^3.5
r^4 «
N
£
k
20 40 60
SWIRL ANGLE, DEGREES
SMOKE > 1.0
SMOKE <_ 1.0
3
2
0
gm
1
0
(
7-90°-B NOZZLE
100 PSIC, 7 CPK
^~ ~
•-BB-K-.
---IT]
-H-B^"^
t
—
.-^--"
^
HOKE
DIA.
4.0
Li.u
V/'-s
^
i
3 20 40 60
SWIRL ANCLE, DECREES
Figure HI- Versatile burner flue gas emissions obtained with 10-percent
excess air in the 30-inch-diameter hot-wall chamber, 7
-------
swirl angle of 25 degrees appear to be generally near optimum at this
flowrate. Note that, at this 7-gph flowrate, the 1 gm NO/kg fuel mini-
mum appears still to be applicable for refractory wall (i.e., near
adiabatic) combustors.
An 18-inch-diameter coaxial chamber, internally insulated with 1/2 inch
of Pyroflex, was used for 3-gph versatile burner tests. The small ver-
sion of the versatile burner was used for these experiments. The
results are shown in Fig. 112. Here again, a 25-degree swirl angle
appears to be a good compromise, while the 2.5-inch choke diameter
is the better of the two shown.
At the 0.75-gph level, versatile burner experiments were conducted
in an 8-inch-diameter, Pyroflex-lined coaxial combustion chamber.
For those experiments, the air swirler vanes were relatively small
(1/2-inch radial dimension by 2 inch length) and ineffective (placed
at the periphery of the blast tube) so the only major geometry variable
was the choke diameter. Shown in Fig. 113 is a spread of NO emission
data obtained as a function of choke diameter. The data spread is for
various nozzle.recesses. The optimum choke diameter appears to be at
about 1.5 inches. These were the first versatile geometry burner
tests conducted and, since the experimental results indicated no
swirler angle effects, all later versatile burner tests were conducted
with swirler blades extending radially inward to within about 0.3 inch
from the oil nozzle. Although no effective air swirl was tested at
0.75 gph, the data trends at higher flowrates suggest that an air
swirler angle of 25 degrees might be most acceptable at this 0.75-gph
level also.
Throughout the 0.75- to 12-gph range, the use of about a 25-degree air
swirler angle appears to be a good compromise to eliminate smoke with-
out inducing high NO emission. However, the optimum, or best compro-
mise choke diameter is apparently dependent on the oil flowrate. The
nominal best compromise choke diameters are shown in Fig. 114 as a
200
-------
2.5
2.0
•8
I
0)
2
o 1.5
I
1
u
•H
1.0
0.5
18-INCH DIAMETER, HOT-WALL
COAXIAL COMBUSTOR
OIL NOZZLE: 3.0-60 -A
NOM. S.R. = 1.30
— —— BACHARACH SMOKE £ 1
BACHARACH SMOKE < 1
I
20" 30"
Air Swirl Vane Angle
40"
50U
Figure 112, Comparison of nitric oxide emission results from the
1- to 3-gph versatile burner experiments as a function
of air swirl vane angle
201
-------
N>
O
to
4.0
•o
I 3.0
O OQ
•H i-l
£3
«fc 2.0
1.0
1.0
8-INCH DIAMETER, HOT-WALL
COAXIAL COMBUSTOR
OIL NOZZLE: 0.75-90 -A
SWIRL ANGLE: 60°
DATA BANDWIDTH
FOR NOZZLE RECESS
VARIATION OF 1/8 IN.
TO 5/8 IN.
I
1.5
2.0 2.5
Choke Diameter, Inches
3.0
3.5
Figure 113. Nitric oxide emissions as a function of choke diameter for the
versatile research oil burner at a nominal oil flowrate of 0.75
gph with 25-percent excess air
-------
to
10.0
3.0 -
0.6
1.0
2.0 4.0
Oil Flowrate, Gallons/Hour
6.0
8.0 10.0
20.0
Figure 114. Choke diameter versus oil flowrate for minimum nitric oxide emissions,
low smoke, and carbon monoxide emissions (applicable to burners
incorporating a six-vane, 25-degree air swirler)
-------
function of oil flowrate. The assessment of the compromise is at best
only semiquantitative and, therefore, the optimum diameter ranges are
shown as shaded area's. A single, straight line appears to adequately
represent the center of the ranges. The equation of that line is:
DIA = (2.7 x gph)0'4
where
DIA = optimum choke diameter, inches
gph = oil flowrate, gal/hr
If an oil burner is designed with the choke diameter obtained from the
above equation, and if a six-vane set of 25-degree air swirler blades
is used in the end of the blast tube, a near-optimum compromise between
smoke and nitric oxide emissions should be expected when the burner is
fired into an appropriately sized coaxial combustion chamber.
All of the versatile geometry burner experiments described above were
conducted in refractory-lined combustion chambers which had their axes
oriented coaxially with the burner blast tube axes. Perhaps a more
common configuration in the 1-gph, residential burner size is the per-
pendicular orientation. For the perpendicular orientation, the burner
blast tube axis is directed radially into the combustion chamber, near
its closed end. The perpendicular orientation requires that flow from
the blast tube make a right-angle turn after entering the combustion
chamber. The right-angle turn tends to disrupt smooth flow patterns,
and probably makes the flow resemble a well-stirred reactor which was
previously shown to promote the formation of nitric oxide. The use of a
perpendicular port combustion chamber is, therefore, expected to be
detrimental to the objective of minimizing nitric oxide emissions; how-
ever, since such a large portion of the burners in the 1-gph, residen-
tial size range utilize the perpendicular port combustion chamber, a
series of experiments was conducted with the 1-gph versatile burner
204
-------
mounted in the perpendicular port combustion chamber. The results of
those experiments are shown in Fig. 115. It is observed that there is
a considerably greater tendency for smoke formation in the perpendicu-
lar port combustion chamber and, also, a higher level of nitric oxide
emissions. The smoke probably results from oil impingement on the
chamber walls, and the high nitric oxide emissions from the flow dis-
ruptions which "stir" the combustion process. For the data in Fig.
115, the 1.0-90°-A nozzle, a 1.5-inch-diameter choke and about a
25- to 30-degree swirler angle appear to be near optimum. The choke
diameter and swirler angle values are very close to the optimum values
previously shown for coaxial combustion chambers. In conclusion, it
was very apparent that the perpendicular port combustion chamber re-
sulted in much worse pollutant formation than coaxial chambers.
205
-------
2.5
ca
u 2.0
£
w
a
u
1.0
S.R. w 1.25
1.0-90 -A OIL NOZZLE
8-INCH DIAMETER, HOT WALL,
PERPENDICULAR PORT COMBUSTOR
BACHARACH SMOKE =£ 1
BACHARACH SMOKE < 1
0°
2.5 r
10° 20° 30°
AIR SWIRLER VANE ANGLE
40°
2.0
NITRIC OXIDE, gm
CHOKE
DIA.
1.0 IN
•isSJN.-
2.0 IN
1.0-60°-A OIL NOZZLE
8-INCH DIAMETER, HOT WALL,
PERPENDICULAR PORT COMBUSTOR
---- BACHARACH SMOKE > 1
BACHARACH SMOKE < 1
0°
10° 20° 30°
AIR SWIRLER VANE ANGLE
40°
2.5 IN.
2.0 IN.
1.5 IN.
50°
50°
Figure 115.
Nitric oxide emission results from 1-gph versatile
burner experiments in a perpendicular port combustor
206
-------
SIMULATED FIELD TESTING OF OPTIMUM GEOMETRY BURNERS
To verify the optimum burner geometry design criteria described in the
previous section, two fixed geometry burners were fabricated according
to those criteria, and subjected to an extensive series of tests. The
two burners were designed for 1 and 9 gph, respectively. The 1-gph de-
sign represented an "extrapolation" since the 1-gph versatile geometry
results did not include the use of effective air swirler blades, and it
was assumed that the 25-degree air swirler angle found to be optimum for
the larger 3-, 7-, and 12-gph size burners would also be optimum for the
1-gph size. The 9-gph design represented an "interpolation"of the ver-
satile geometry results, which were conducted at 1, 3, 7, and 12 gph.
1-gph Fixed Geometry Optimum Burner
The 1-gph fixed geometry burner based on the optimum design criteria,
included a 1.5-inch choke ring diameter and 25-degree air swirler vanes
installed in a 4-inch-diameter blast tube. The air swirler blades ex-
tended inward to the 0.6-inch radius and were 2.06 inches long in the
axial direction. A photograph of this burner is shown in Fig. 116.
Test results obtained with this burner are shown in Fig. 117, indicating
that it did achieve the desired result of 1 gm NO/kg fuel without smoke.
Also shown in Fig. 117 is the range of NO of emissions experienced using
several commercially available burners under similar conditions. These
results indicate that the use of the optimum geometry criteria described
above do provide a definite reduction of NO compared to the typical com-
mercially available burners for smoke-free conditions. Shown in Fig. 118
are optimum burner and commercial burner data obtained using a cylin-
drical combustor with the burner blast tube port perpendicular to the
axis of the chamber. The NO levels with the perpendicular port chamber
are significantly higher than with the coaxial chamber, apparently be-
cause of the flow disruptions (i.e., combustion zone stirring) induced
by the geometry change. However, note that under low smoke conditions,
the optimum geometry burner is still lower in NO than the typical
207
-------
o
CO
',/ /
5DZ21-8/6/73-S1A
Figure 116. Closeup photograph of the 1-gph optimum oil burner head
-------
NJ
o
-------
10
(-1
o
3.5
3.0
-------
commercial burners, even though the optimum was determined from coaxial
combustion chamber studies. Comparison of Fig. 117 and 118 supports the
desirability of plug-flow-type combustion (which apparently cannot be
achieved in a perpendicular part combustor due to flow disruption in the
required right-angle turn) and the superiority of combustion chambers
which are oriented coaxially to the burner blast tube.
Perpendicular port combustion chambers are very common in the 0.5- to
1.5-gph home heating applications where minimum floor space is available
for the furnace. It is unfortunate that the very common perpendicular
port geometry generally leads to higher NO emissions (by disrupting the
flow, making the combustor resemble a near-stiochiometric well-stirred
combustor rather than a near-stoichiometric, plug-flow combustor). The
lower NO levels characteristi-c of coaxial chambers should be an incentive
to alter furnace designs to allow universal use of coaxial combustion
chambers.
A total of 128 hours of cyclic testing time (42 hours burning time) was
accumulated as a part of simulated field testing with this burner in the
Unitron A100 hydronic/warm-air furnace. The cycles were 5 minutes on,
10 minutes off, 5 minutes on, 10 minutes off. Nitric oxide, unburned
hydrocarbon, and carbon monoxide emissions as a function of time are
shown in Fig. 119. The high carbon monoxide and unburned hydrocarbons
at the termination of the firing cycle are not extremely significant
since the analysis system was sampling stagnant gas from the stack. The
performance of the 1-gph optimum burner over the entire test period
showed little variation and no noticeable degradation. The nominal
stoichiometric ratio, with no mechanical readjustment, drifted upward
slightly, ~2 percent (from about 1.08 to 1.10 to 1.10 to 1.12). This
was probably due to changes in ambient conditions associated with the
outdoor test facility being used. None of the exhaust pollutant emis-
sion concentrations increased during the testing, with nitric oxide at
about 0.95 gm NO/kg fuel burned, no measurable unburned hydrocarbons,
no smoke (Bacharach smoke scale = 0), while carbon monoxide improved
211
-------
in
1
u
•O ft,
Q)
•33
•HU,
0«H
x
O
rt
u
25
20
IS
10
5
0
1.0
0.5
5.0
4.0
1.0
0
u
I
Time, Minutes
Figure 119. Exhaust gas composition profiles as a function
of time for the 1-gph optimum oil burner in a
refractory-lined chamber, hydronic furnace
212
-------
(0.60 •*• 0.35 gm CO/kg fuel) probably due to the stiochiometric ratio
shift. Removal and inspection of the 1-gph optimum burner after comple-
tion of testing revealed no soot or sludge accumulation nor any evidence
of deterioration of either the burner or the furnace.
In summary, the 1-gph fixed geometry burner behaved as well, or better,
than expected in all respects.
9-gph Fixed Geometry Optimum Burner
The 9-gph fixed geometry optimum burner blast tube was mounted on the
Sun-Ray burner body. It had a 3.75-inch-diameter choke ring installed
in a 6-inch-diameter blast tube, with six 25-degree air swirler vanes
extending into the 0.6-inch radius, and 2.8 inches long in the axial
direction. This optimum burner, with a 9.0-45°-B oil nozzle, was fired
in the 30-inch-diameter, water-cooled, coaxial combustor. The resultant
nitric oxide emission as a function of stoichiometric ratio is shown in
Fig. 120 along with corresponding profiles from the 12-gph Sun-Ray and
the 6-gph Nu-Way burners. The differences are quite substantial, the
9-gph optimum burner producing clearly 50-percent lower nitric oxide
emissions in the smoke-free operating range than either of the other
burners. The 9-gph optimum burner operated smoke-free to near-
stoichiometric conditions (SR —1.02), and at a nominal nitric oxide
emission level of about 0.65 gm NO/ kg fuel burned. This level of NO
emissions for the 9-gph optimum burner is very low, relative to the
1-gph optimum burner data shown in Fig. 117 (NO « 1.10 gm NO/kg fuel)
primarily because of the effect of the cold-wall (Twall « 400 F) com-
bustor and probably due in part to the large combustor volume.
The 9-gph fixed geometry optimum burner consistently ran somewhat noisier
than the conventional burners in the 10-foot-long, 30-inch-diameter com-
bustor, whereas the 1-gph fixed geometry burner was noisy only under
rare circumstances. Whether the noise was being generated as a result
of combustor resonance and/or as a result of flame instability was not
213
-------
2.0
o
Ul
u
u.
1.5
1.0
§
>—i
X
o
S 0.5
30-INCH DIAMETER, WATER-COOLED,
COAXIAL COMBUSTOR
BACHARACH SMOKE >. I
BACHARACH SMOKE < 1
SUN-RAY 12 GPH
OPTIMUM BURNER
9 GPH
I
I
I
1.00
1.10 1.20
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
Figure 120. Comparison of nitric oxide emission profiles for various oil burners in
the 6- to 12-gph firing rate range
1.30
-------
apparent. A high sensitivity microphone was used to measure the noise
being generated by the 9-gph fixed geometry optimum burner in the 30-
inch- diameter, cold-wall combustor. Figure 121 shows a photograph of
an oscilloscope display of the noise (microphone output) obtained during
the burner firing. These data indicate that the frequency of the noise
is about 39 to 40 Hz. Assuming an average temperature in the chamber
of 2000 R, the acoustic velocity was calculated to be about 2100 ft/sec.
The estimated effective length of this combustor has a quarterwave reso-
nance which corresponds to a resonant frequency of 43 Hz, which is reas-
onably compatible with the measured frequency. Therefore, it appears
likely that the noise is due to a quarterwave resonance in the combustor.
This being the case, it is likely that if the burner were installed in
a combustor of smaller volume, the resonance would disappear and the
burner would run quietly. In practical commercial applications, the use
of combustors of smaller volume than the one used here is standard
practice.
Simulated field testing of the 9-gph optimum burner also was conducted
with a total cycle time of 112 hours (27 hours of burning time) being
accumulated with this burner in the 30-inch-diameter, water-cooled,
coaxial combustor. The most usual cycle was 5 minutes on, 10 minutes
off. The performance of this larger, optimum burner also showed little
variation and no noticeable degradation. The stoichiometric ratio in-
creased slightly from 1.02 to 1.03 to 1.03 to 1.04 during the test
series. A slight improvement was noted in the NO emission level which
decreased from about 0.5 to 0.6 gm NO/kg fuel to 0.45 to 0.50 gm NO/kg
fuel burned. The CO emission level varied slightly, probably with chan-
ges in ambient conditions, but CO emissions as low as 0.6 to 0.7 gm CO/
kg fuel burned were still being obtained at the end of the test series.
The unbumed hydrocarbon emission level was zero, and the Bacharach
smoke level was zero. Time-dependent hydrocarbon and nitric oxide emis-
sions for a typical cycle are shown in Fig. 122. Inspection of the 9-gph
optimum burner at the end of the test series revealed only a slight oily
accumulation around the perimeter of the 5-inch-diameter burner head,
215
-------
40 cyc/sec
0.1 0,2 0.3
Time, seconds -
0.4
Figure 121. Oscilloscope frequency trace of noise generated
by the 9-gph fixed geometry optimum burner,
measured at the exhaust of the 30-inch-diameter,
10-foot-long coaxial combustor
216
-------
o
•B
rt
o
I
•o
0)
5.
a
3
u.
u
Si
i>
25
20
15
10
5
0
0.5
rt
+J
(O
o
•o
4-1
CO
Figure 122. Exhaust gas composition profiles as a function of
time for the 9-gph optimum oil burner in a cold-
wall, 30-inch-diameter, coaxial combustor
217
-------
and no evidence of deterioration of the burner head assembly. An inter-
esting characteristic of this burner is the almost invisible flame it
produced, implying an extremely low particulate concentration in the
flame, and a high degree of recirculation.
This optimum geometry burner also behaved as well as, or better than,
expected in all respects.
NONCONVENTIONAL BURNERS
The primary emphasis in this program was on relatively conventional,
high-pressure atomizing, luminous flame, No. 2 distillate oil burners
fired into refractory combustion chambers. The objective was to make
minor changes in the burner blast tube end which would result in a re-
duction in emissions and improvement in efficiency. The changes were
desired to be such that they could be retrofitted onto existing burners
and they should not introduce any new or special servicing requirements.
In addition to the conventional burner development, a portion of the
program effort was set aside for application to nonconventional burners.
These nonconventional burner concepts were generally felt to be poten-
tially applicable to future burners, but cost and serviceability were
not a significant consideration in selection of the concepts to be
tested. Several different nonconventional burners were tested, includ-
ing the intense swirl burner, the forced combustion gas recirculation
burner, the displaced injection burner, the two-stage burner, the three-
stage burner, and the nonswirl optimum geometry burner. These burners
and the results obtained with each are described in the following
sections.
Intense Swirl Burner
Results obtained with the commercial burners suggested that a well-
stirred combustion process would produce relatively small amounts of
>
nitric oxide when operated at high excess air levels. To verify this,
218
-------
an existing 0.5- to 3-gph commercial burner was modified to accept a
variable rate, mechanically rotated, multiple-vane swirler assembly to
be used to investigate low to very high air swirl rates. This intense
swirl burner has been described earlier, in the Apparatus section of
this report. Figures 23 and 25 in that section show a schematic and
a photograph of the modified burner with the motorized swirler assembly.
The high-swirl burner was tested with maximum (3450 rpm, swirl No.=8.8)
swirl, with low (812 rpm swirl No.=2.2) swirl, and with no swirl. The
burner was fired in the 8-inch-diameter, perpendicular port, refractory-
lined combustion chamber. The nitric oxide emissions determined from
analysis of the mixed exhaust gases are shown in Fig. 123. When there
was no swirl, the nitric oxide emissions were similar to those (observed
with typical commercial burners, except there was a very noticeable lack
of repeatability. The lack of repeatability was later determined to be
due to a nonstationary flame front location. Apparently, the smooth
convergence of the choke device, combined with the lack of any swirl at
all, and the absence of a continuous spark igniter, allowed the flame
front to shift from test to test.
At low swirl, the nitric oxide emissions were moderately high, and var-
ied irregularly with stoichiometric ratio.
At very high swirl rates, the nitric oxide emissions decreased monotonic-
ally with increasing stoiochiometric ratio. This is the well-stirred
reactor trend which was expected from interpretation of results from the
commercial burner studies, as discussed in pages 173 through 188. The
well-stirred burner obviosuly has the potential for extremely low nitric
oxide emissions at high excess air, especially considering the fact that
these results were obtained in a refractory-lined chamber; however, the
furnace efficiency penalties (due to lower temperatures for heat trans-
fer) which result from operations at high excess air are not tolerable.
219
-------
2.5
2.0
g
u 1.0
8
M
0.5
8.0" OIA. CYLINDRICAL COMBUSTOR
90 BURNER PORT
REFRACTORY LINED CHAMBER
INTENSE SWIRL
(3450 RPH)
BACHARACH SMOKE < 1
BACHARACH SMOKE > 1
\
1.0 1.5 2.0 2.5
STOICHIOMETRIC RATIO. (AIR/FUEL)/14.49
3.0
Figure 123. Nitric oxide emissions, comparison of the effect of
swirl on the 30-degree convergence, 1.65-inch-
diameter, mechanically rotated, six-vane swirler,
with no spark igniter
220
-------
If this same burner were fired at high swirl in a water-cooled combustor,
the nitric oxide levels would undoubtedly be much lower, and the enhance-
ment of heat transfer to the combustion chamber walls by virtue of the
high swirl velocities might wholly compensate for the penalties of high
excess air operation.
Forced Combustion Gas Recirculation Burner
The benefits of flue gas recirculation for the control of nitric oxide
emissions are well known. However, it is not convenient to arrange for
flue gas recirculation in typical domestic applications. A special bur-
ner was fabricated to achieve the benefits of flue gas recirculation
without any modifications to the heating plant, other than the burner
itself. This burner is constructed to suck combustion gases out of the
combustion chamber and mix them with the incoming air. The burner was
designed to operate in a cooled combustion chamber, so that the combus-
tion gases withdrawn by the burner would be at least partially cooled
to yield the same benefits as flue gas recirculation. The flow direc-
tions are shown schematically in Fig. 27, in the Apparatus section,
and a photograph of the assembled burner is shown in Fig. 28.
Tests were conducted with this burner, using an 0.5-gph spray nozzle,
varying the overall stoichiometric ratio and varying the ratio of recir-
culated combustion gas to primary air in the burner. The overall stoi-
chiometric ratio was determined by chemical analysis of the flue gas
composition. The recirculation ratio in the burner was determined by
measuring the composition of the mixed gases in the burner blast tube.
Since, in the analysis of the mixed gases, it was not possible to dif-
ferentiate between oxygen coming in with the fresh air and unburned
i
oxygen coming in with the recirculated gas, the recirculation ratio is
reported as the weight ratio of burned gases to unburned gases. For
this definition, the burned gases include CO, C02> H_0, and NO, as well
as the inert atmospheric gas (NO associated with the oxygen in the com-
bustion products; the unburned gases are the free oxygen and the inert
221
-------
atmospheric gas associated with it. This recirculation ratio can be
calculated from 0_, CO, and C0» concentrations in the mixed gases accord-
ing to the following formula:
Recirculation Ratio =
Y v 7Q Y 1 Y 7Q
C02 [44 + | 18 + (1 + J) -g- 28] + CO [28 + J 18 + (i + ±) li 28]
°2 (32 + IT 28)
where
recirculation
ratio = grams of burned gas per gram of unburned air
C0_ = volume percent CO in the mixed gases, dry basis
CO = volume percent CO in the mixed gases, dry basis
0 = volume percent 0» in the mixed gases, dry basis
x = hydrogen to carbon atomic ratio for the fuel, x = 1.814
The forced combustion gas recirculation burner was tested in the 8-inch-
diameter coaxial combustion chamber. For these tests, the coaxial com-
bustion chamber did not have a refractory lining, and the exterior of
the steel combustion chamber was cooled by natural air convection. The
flame produced by this burner was mostly nonluminous, with a few streaks
of blue. The nitric oxide emissions measured for these tests are shown
in Fig. 124. This burner produces very low nitric oxide emissions, even
at near-unity stoichiometric ratios. A recirculation ratio of 0.16 grams
burned gas per gram of unburned air appears to be about optimum for this
burner. The highest recirculation ratio of 0.4 produced slightly more
nitric oxide than the ratio of 0.16, presumably because the gases recir-
culated at the higher ratio were apparently drawn from the uncooled core
of the combustion process whereas those at a ratio of 0.16 were mainly
from the cooler regions near the wall. This presumption is supported
by the observation of a higher carbon monoxide-to-burned fuel ratio for
the high recirculation ratio cases.
222
-------
g
I
u.
ti.
O
g
X
O
1.0
0.8
0.6
0.4
z 0.2
\
8 IN. DIA.. AIR-COOLED COMBUSTOR
OIL NOZZLE: O.SO-80°-C
BACHARACH SMOKE & 1
BACHARACH SMOKE < 1
RECIRC RATIO*
0.16
0.40
*RECIRC. RATIO
LBS OF BURNED GAS
LB UNBURNED AIR
0.90 1.00 1.10 1.20 1.30 1.40
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
1.50
1.60
Figure 124. Effects of varying recirculated combustion gas with
the forced combustion-gas recirculation burner
-------
It is notable that the nitric oxide emissions for this 0.5-gph burner
are a factor of 2 or 3 lower than for the 1-gph optimum geometry burner.
This is due to: (1) operation in a cool wall chamber (as opposed to a
refractory wall chamber for the optimum burner, and (2) the recircula-
tion of cooled combustion gas products through the burner, where the
cooled gases lose even more heat through the walls of the burner to the
surrounding environment.
An additional factor of interest in Fig. 124 is the tendency for this
burner to operate smoke-free. At the high recirculation ratio of 0.4,
the burner operated smoke free even in the fuel-rich region to the left
of unity stoichiometric ratio in Fig. 124. Under this condition, there
was, of course, a considerable amount of carbon monoxide, but very low
nitric oxide and smoke emissions. The tendency for smoke-free operation
appears to be enhanced by the higher recirculation ratios, suggesting
that the temperature of the mixed primary gas is a strong factor. Appar-
ently the hotter primary gas tends to prevaporize the fuel, prior to the
start of combustion, reducing the tendency for smoke formation. This
factor may also explain the general flame appearance, which was largely
nonluminous with a few streaks of blue.
Using the same combustion chamber, but installing a 1/2-inch-thick Pyro-
flex liner, the burner was fired in the hot-wall configuration. The
results of these tests are shown in Fig. 125, compared to some of the
previously presented cool-wall results. The results with the hot-wall
configuration are essentially as might be expected from consideration
of the fact that the recirculated combustion gases are hotter with the
hot-wall chamber. The nitric oxide emissions are considerably higher,
undoubtedly an effect of the higher temperatures. The smoke-free region
also appears to be wider, apparently due to the greater tendency for pre-
vaporization of the fuel.
224
-------
1.0 _
in
a
iu
0.8
u.
o
g)
g
o
0.6
- I
0.2
0.8
\
45 (HOT WALL)
0.40 (COOL WALL)
I
0.9
Figure 125.
8 IN. DIA., COMBUSTOR
OIL NOZZLE: O.SO-80°-C
BACHARACH SMOKE & 1
BACHARACH SMOKE < 1
*RECIRC RATIO
LBS OF BURNED GAS
LB OF UNBURNED AIR
RECIRC
RATIO*
0.1S (HOT WALL)
0.16 (COOL WALL)
1.0 1.1 1.2 1.3
STOICHIOMETRIC RATIO. (AIR/FUEL)/14.49
1.4
1.5
Effects of combustion chamber wall temperature with
forced combustion-gas recirculation burner
-------
Tests were also conducted with a larger flowrate spray nozzle in the
forced combustion gas recirculation burner. The results of these tests
are shown in Fig. 126. To achieve recirculation ratios above 0.08 gram
of burned gas per gram of burned air required.the use of a lesser number
of lower-angle swirler vanes (because the blower power was limited and
the swirler vanes were the main pressure drop). Note that, in general,
the 0.75-gph flowrate resulted in higher nitric oxide emissions, pre-
sumably because of the higher heat loads at 0.75 gph and the higher
rates of recirculated gas flowrate combining to cause higher mixed gas
temperatures (but perhaps affected by the different swirler vane config-
uration). The higher mixed gas temperatures are less effective in reduc-
ing nitric oxide formation. The forced combustion gas recirculation bur-
ner nitric oxide emissions at 0.75 gph would very likely be improved by
use of combustion gas entry ports (Fig. 27) located at a larger radius
where they would have a greater tendency to pull in combustion gases
from the vicinity of the cooled walls.
In conclusion, burners incorporating forced combustion gas recirculation
appear to offer the combined benefits of smoke-free operation at very
near to stoichiometric air/fuel ratios and low nitric oxide emissions.
They must, however, be operated in conjunction with air-cooled or water-
cooled combustion chambers. Some additional advantage might also be ob-
tained if the combustion chamber geometry were specially tailored to
enhance recirculation of gases to the burner end of the chamber.
Displaced Injection Burner
The displaced injection burner concept is shown schematically in Fig. 26
for two configurations that were tested. The burner design is intended
to take advantage of combustion gas circulation induced by tangential
injection of the combustion air. As shown in Fig. 26 (Apparatus section),
the tangential injection of combustion air imparts a strong circulation
to the gases, causing combustion gases to be intermixed with the fresh
air. The oil injection nozzle is located so that the oil is sprayed into
226
-------
to
1.0..
0.8.
« o.
o
i-t
X
o
RECIRC RATIO'
0.9
1.0
0.08 (8-VANE, 60° SWIRLER)
0.16 (4-VANE. 25° SWIRLER)
0.16 (NO SWIRLER)
0.20 (4-VANE/ 25" SWIRLER)
8-INCH DIAMETER, AIR-COOLED, COAXIAL
COMBUSTION CHAMBER
0.75-80°C OIL NOZZLE
*RECIRC RATIO = "S OF BURNED GAS
LB OF UNBURNED AIR
BACHARACH SMOKE >, 1.0
BACHARACH SMOKE < 1.0
I
1.5
1.6
Figure
1.1 1.2 1.3 1.4
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
126. Forced combustion gas recirculation burner emissions at an oil flowrate of 0.75 gph
-------
a mixture of burned gases and fresh air. The recirculation pattern
induced by the burner was intended to take advantage of the combustion
gas cooling which results as the gases pass along the wall. Thus, the
design shown in Fig. 26 was intended to achieve the same results as the
forced combustion gas recirculation burner described in the .previous sec-
tion, but without requiring the passage of combustion gas through the
air fan.
The displaced injection burner was tried only in the two configurations
shown in Fig. 26 and, for each of these, it resulted in relatively heavy
smoke emissions, even though it was tested in a refractory-lined chamber.
The smoke apparently was caused by impingement of oil spray on the com-
bustor walls, and/or poor mixing of the air/oil spray mixture. Further
studies of this burner concept, employing a larger combustion chamber,
and/or variations in the oil spray injection site, would be worthwhile.
Two-Stage Burner
If combustion takes place under fuel-rich conditions and chemical equi-
librium is achieved, there is a tendency for relatively low nitric oxide
emissions. This may or may not be offset because of superequilibrium
amounts of nitric oxide formed by the prompt reaction mechanisms. Most
oil burners operate with all of the combustion taking place under air-
rich conditions where nitric oxide chemical equilibrium is unfavorable
to the objective of low nitric oxide emissions. A burner (Fig. 127)
was designed that carried out all but the last stage of combustion under
fuel-rich conditions.
It was hoped that the fuel-rich portion of the combustion could be car-
ried out smoke free with the use of intensely swirled air (3450 rpm).
The grounds for this optimism was the smoke-free nature of fuel-rich
combustion in the forced combustion gas recirculation burner, and with
the intense swirl burner. However, preliminary tests with the burner
shown in Fig. 127 did not result in smoke-free operation, so further
testing was not conducted.
228
-------
NJ
N)
By-Pass Secondary Air
First Stage
Fuel Rich
Combustion
By-Pass Secondary Air
Figure 127. Schematic of the two-stage, intense-swirl, concentric-
tube oil burner head extension
-------
Three-Stage Burner
As a further effort to achieve smoke-free, fuel-rich combustion, a three-
stage burner was conceived. The three stages of combustion are: (1) pre-
heat the inlet air to 400 to 600 F, (2) use the air for fuel-rich, low
nitric oxide, hopefully smoke-free combustion, and (3) add the last bit
of air to make the overall combustion air rich, thus eliminating carbon
monoxide and unburned hydrocarbons with a minimal increase in nitric
oxide. The test configuration for this burner is shown in Fig. 29
(Apparatus section). The test configuration included only the first two
stages of the three-stage device. The device shown in Fig. 29 was
tested to determine whether or not it could burn fuel rich with no smoke
emissions. The results from these tests are shown in Fig. 128, where it
is noted that the nitric oxide emissions are very low but, in all cases,
the Bacharach smoke level exceeded a reading of 1. The smoke readings
ranged from 2 to greater than 9 in the fuel-rich region. Since smoke-
free operation was not achieved in the first stage, testing of the remain-
ing two stages of combustion was not conducted.
No-swirl Optimum Geometry Burner
The earliest burner geometry optimization experiments utilized a versa-
tile burner that had relatively ineffective swirler vanes and, hence, a
fixed geometry burner was constructed that had a choke diameter corre-
sponding nominally to the optimum geometry criterion (dia. = [2.7 gph]4)
but which had no swirler vanes. In a further attempt to streamline the
flow, the burner was built, as shown in Fig. 129, with the choke in the
form of a smoothly converging and diverging contour. The diverging part
of the contour was made removable so that the burner could be tested
with only the converging section.
Several tests were conducted with this burner, a few of the results of
which are shown in Fig. 130. The tests were conducted with two oil spray
nozzles, and with/without the 45-degree expansion section on the nozzle.
230
-------
1.0
0.8
0.6
u.
o
\
\
\
\
8 IN. INSIDE CHAMBER DIAMETER
OIL NOZZLE: 0.50-80°-C
2 IN. DIAMETER, 75° SWIRL RING
BACHARACH SMOKE
BACHARACH SMOKE
\
\
\
•AMBIENT AIR
8-VANE SWIRLER
. 0.4
w
Q
H^
X
o
u
M
£
M
z 0.2
0
^*"* ~ *** ** "^
^^
^^* ^^ ^^ ^^" ^^* ^^^^
^. — — — ^V PREHEATED AIR -
>^ 8-VANE SWIRLER ^^. *-^
. \ -^ V
— •
• •!•••• 1 ••
0.90 0.95
.,
^^^^
^
^^^
^^
*^
PREHEATED AIR
4-VANE SWIRLER
1
1.00
STOICHIOMETRIC RATIO,.(AIR/FUEL)/14.49
Figure 128. Nitric oxide emission results of fuel-rich combustion experiments
with a fixed-vane intense-swirl burner head
-------
Removable
Expansion
Section
30° Convergent
Section
ro
w
to
Figure 129. Cross-section view of the fixed geometry, 1.65-
inch-diameter no-swirl optimum burner head
-------
2.5 _
2.0
s
H.
U.
o
H 1'°
O
t-i
X
o
g
z
O.S
1.0
8.0" DIA. CYLINDRICAL COAXIAL
HOT WALL COMBUSTION CHAMBER
BACHARACH SMOKE < 1
— BACHARACH SMOKE ^ 1
30° CONV. + 45° EXP.
75-70°-A
30^
0.75
CONV.
-70°-A
-30 CONV.
0.75-90°-A
1.5 2.0 2.5
STOICHIOMETRIC RATIO, (A/F)/14.49
3.0
Figure 130. Nitric oxide emissions of the fixed geometry,
convergent and expansion sections head, optimum
geometry, no-swirl burner
233
-------
Operation with the 90-degree nozzle and without the 45-degree expansion
section was found to be best, as shown in Fig. 130. However, later
attempts to repeat the low nitric oxide emission data resulted in a wide
data spread, as shown in Fig. 131.
Because of the smooth flow profiles that should be expected from the
smoothly contoured choke, it was suspected that the poor data repeata-
bility was due to poor stability of the flame front location. A series
of tests was conducted with an annular-shaped, flame-holding ring, and
this seemed to stabilize the data. It was further suspected that the
spark ignition device was influencing the flame front location, and so
a series of experiments, shown in Fig. 132, was conducted (no flame
ring), varying the location and condition of the igniter. As shown in
Fig. 133, the nitric oxide emission level was lowest and most repeatable
when the igniter was either turned off or extended downstream of the
spray nozzle. This suggests that the flame front location is optimum
when it is relatively far downstream of the spray nozzle.
The difficulty with data reproducibility resulting in the occasional
relatively high nitric oxide emissions was apparently the combined effect
of the smooth, no-swirl flow, and test-to-test variation in the effective-
ness of the igniter. The fixed geometry optimum burner.test results,
reported in a previous section of this report, did not have similar data
reproducibility problems, apparently because those burners incorporated
six 25-degree air swirler vanes which tended to stabilize the flame front
location.
To determine if the spark igniter had a similar effect on the action of
other burners, tests were conducted with the ABC 55-J burner, and a EPA-
furnished version of the ABC Mite burner (similar, but not identical to
the Mite burner described in other parts of this report). The results of
those tests are shown in Fig. 132, where it is apparent that, in some
cases, turning the igniter off improves nitric oxide emissions while, in
others, it increases them.
234
-------
Q
w
s
u.
o
g
l-l
g
u
2.0__
1.5
1.0
8.0" DIA. COAXIAL CYLINDRICAL COMBUSTOR
L
— BACHARACH SMOKE < 1
BACHARACH SMOKE > 1
SEPT. 21, 1972
..— SEPT. 15, 1972
- OCT. 17, 1972
OCT. 4, 1972
0.5
AUG. 22, 1972
1.0
1.5 2.0 2.5
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
3.0
Figure 131. Variation of the nitric oxide profiles for the
1.65-inch-diameter, fixed geometry versatile
burner with a 0.75-90°-A nozzle tested on dif-
ferent days
235
-------
•S
o
S
•H
2.0
« 1.5
I
1.0
0.5
8.0" DIA. COAXIAL CYLINDRICAL
HOT WALL COMBUSTOR
No Igniter
Std. Igniter
— BACHARACH SMOKE < 1
BACHARACH SMOKE > 1
1 ABC Mite
J (E.P.A.)
Std. Igniter
No Igniter
ABC 55-J
Extended § No-Igniter
1.65" Oia. Fixed Geom.
Smooth Contour,
No-Swirl Burner
I
1.0 1.5 2.0 2.5
Stoichiometric Ratio, (Air/Fuel)/14.49
Figure 132. Comparison of nitric oxide emissions of com-
mercially available burners with spark igniter
modifications
236
-------
1.5
tu
o
1.0
Ul
o
l-l
x
o
u
M
^ 0.5
2
8.0" DIA. CYLINDRICAL COMBUSTOR
BACHARACH SMOKE < 1
BACHARACH SMOKE > 1
STANDARD IGNITER
DATA SPREAD
LAST RE-RUN OF STANDARD
IGNITER CONFIGURATION
NO IGNITER
EXTENDED IGNITER
• 1.0" FROM NOZZLE
NO IGNITER
1.0 1.5 2.0 2.5
STOICHIOMETRIC RATIO, (AIR/FUEL)/14.49
Figure 133. Comparison of nitric oxide emissions from
the 1.65-inch-diameter, no swirl, fixed geo-
metry, versatile burner with spark igniter
modifications
237
-------
CONCLUSIONS
Several major conclusions can be drawn from the results of this study,
as described in the following paragraphs.
OPTIMUM BURNER GEOMETRY
For high-pressure atomizing, luminous flame burners fired into refractory-
lined combustion chambers, minimum pollutant emissions are obtained with
burners having: (1) no flame-retention device, (2) choke diameters equal
0.4
to (2.7xgph) ' (inches), and (3) peripheral swirler vanes oriented at
25 degrees relative to the blast tube axis. This swirler vane angle
gives the best compromise between smoke emissions and nitric oxide emis-
sions, while the recommended choke diameter produces minimum nitric oxide
emisions.
BURNER DESIGN PHILOSOPHY
With oil burners fired into adiabatic combustion chambers (i.e., refrac-
tory lined chambers) at low excess air levels, formation of nitric oxide
is minimized by the absence of strong mixing and recirculation and the
absence of steep gradients in the combustion zone. That is, with low
excess air and a refractory chamber, a uniform dispersion of fuel spray
in air flowing and reacting smoothly down the combustion chamber, in the
manner of a plug flow reactor, produces the least nitric oxide emissions.
On the other hand, a vigorous intermixing of incoming fresh air with
combustion products (as a result of adiabatic internal recirculation or
steep gradients in the combustion field) tends to promote the formation
of nitric oxide.
238
-------
For adiabatic combustion chambers, the minimization of vigorous mixing
and steep gradients prevents the use of flame-retention devices because
these flame-retention devices promote undesirable mixing and recirculation
in the flame zone. Flame-retention devices are used primarily because
they promote stability of the flame front location so that a particular
burner can be fired into a wide variety of combustion chambers without
special changes or "tuning" of the design for each of the chambers. The
detrimental effects of the flame-retention devices, with respect to pol-
lutant emissions, suggests that they should not be used with adiabatic
combustion chambers and, instead, burners designed according to the opti-
mum burner geometry criteria presented herein should be used, with trial
and error selection of the most suitable oil spray nozzle where necessary.
When oil burners are fired into nonadiabatic combustors (i.e., water-
cooled or air-cooled chambers), the recommended design philosophy is
somewhat different. In this case, external recirculation (i.e., recir-
culation between the flame core and the cool wall) is beneficial, whether
it is produced by the forced recirculation technique described in the
Nonconventional Burners section or by other techniques. In this case,
the recirculated gases are cooled by the cool chamber wall, and when
these cooled gases are mixed with the incoming fresh air, benefits simi-
lar to those obtained with flue gas recirculation, are realized because
of a reduction in peak flame temperature reduction. It appears likely
that internal recirculation (i.e., recirculation within the core of the
flame where there is no opportunity for significant heat loss from the
recirculated gases) is probably most detrimental to nitric oxide emis-
sions, even in the case of nonadiabatic combustion, for essentially the
same reasons it is detrimental in the case of adiabatic combustion cham-
bers. Consequently, flame-retention devices are probably not desirable
in cool-wall chambers either.
239
-------
COMBUSTION CHAMBER WALL TEMPERATURE
Water-cooled or air-cooled combustion chambers characteristically lead
to lower nitric oxide emissions than chambers of similar geometry that
have refractory linings. This is undoubtedly the result of higher flame
temperatures in the refractory-lined chambers. In the cool-wall chambers,
peak flame temperatures are reduced because of heat lost to the chamber
walls by radiation and because of recirculation of cooled combustion
gases, which tends to dilute air injected through the burner blast tube.
The advantages of cool-wall combustion chambers can be exploited through
the use of burner designs that promote the recirculation of cooled com-
bustion gases and the intermixing of these gases with the injected air.
COMBUSTION CHAMBER GEOMETRY
In the case of refractory-lined combustion chambers, it has been shown
that cylindrical, coaxial combustion chambers (i.e., with the chamber
and blast tube aligned along a common axis) result in lower nitric oxide
emissions than cylindrical combustion chambers with the axis of the bur-
ner blast tube perpendicular to that of the chamber. Whether or not a
similar effect occurs with cooled combustion chambers has not been
determined experimentally.
240
-------
APPENDIX A
COMMERCIAL DESIGN PRACTICES
The objective of this portion of the effort was to define present-day
commerical burner design practices. Initially, a listing of oil burner
manufacturers was made from: (1) the Thomas Register, and (2) the Oil
Burner Rating Handbook by Edwin M. Field (printed by Oildum Publications,
Bayonne, N. J.). For each company line, this handbook contains the ca-
pacity, nozzle type, nozzle angle, motor, fuel pump, transformer, controls
and electrode specifications. From this list, companies located through-
out the mid-west and eastern states were contacted by telephone. Bro-
chures of the product line, installation manuals, and service manuals were
requested from each company that manufactured oil burner units in the 0.5-
to 12-gph size range. A list of all companies contacted by telephone is
contained in Table A-l. A high attrition rate of small companies in the
oil burner business is evidenced by the large number of companies that no
longer produce oil burners. (Other companies whose telephone listings
were no longer connected were not included in Table A-l; an additional 17
companies.) Brochures were requested from 26 companies, and 19 were re-
ceived. Without exception, company representatives were interested and
eager to discuss all aspects of their designs and willing to share their
experience and whatever test results they had with us. At each company,
both•engineering and sales personnel were contacted so that some indica-
tion of the business aspects of the company, as well as the type of de-
velopment efforts being conducted, could be determined. Based on the in-
formation obtained from the company brochures and the telephone conversa-
tions, and recommendations made by Dave Locklin of Battelle (a member of
NOFI), five companies were then chosen for plant visits. All of the
241
-------
Table A-l. OIL BURNER MANUFACTURERS CONTACTED BY TELEPHONE
Company Name
Preferred Utilities Mfg.
Victor Mfg. 6 Uistr.
Dunham- Busch Inc.
Carl in Co.
Ace Lngineering Co.
Auto Burner Co.
Croak Engineering
Nu-Way (White-Rogers)
Aldrich Div.
Wayne Home Equip. Co.
Power Flame Div.
General Automatic Prod.
Lynn Products Co.
American Steel Works
American Furnace Co.
Majic Servant Prod.
Calmac
W . N . Bes t Comb . Engr . Co .
Quiet Auto. Burner Co.
Hydrothcrm Inc.
Aeroil Prod. Co.
Acme Heat 6 Power Inc.
Sun Ray
Way-Wolff Assoc.
Reif-Rexoil lac.
Superior Combustion
Cowan Frederick & Co.
Gold Start Oil Burner
State
Location
Connecticut
Illinois
Indiana
Kansas
Maryland
Massachusetts
Missouri
Missouri
Michigan
New Jersey
New York
Remarks
Burners 20 gph only
No longer manufactures oil
burners
Commercial sizes only
Also made plant visit
Commercial sizes only
Also made plant visit
-
Also made plant visit
-
Major Company
-
No longer manufactures
oil burners
No longer manufactures
oil burners
Make heaters for tar only
-
Gas burners only
_
Commercial sizes only
-
-
Commercial sizes only
_
Also made plant visit
0.5 to 8 gph only
Going out of business
0.5 to 225 gph units
International Co.
~
Brochure
Received
No
No
Yes
Yes
No
Yes
No
Yes
No
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
No
Yes
Yes
No
Yes
No
Yes
242
-------
Table A-l. (concluded). OIL BURNER MANUFACTURERS CONTACTED BY TELEPHONE
Company Name
Babcock & Wilcox
Smokeless Oil Burner
North American Mfg.
Pyronics
R. W. Beckett
Bethlehem Corp.
Thermal Research & Engr.
Burnham Corp.
liauck Mfg. Co.
Orr 6 Sembower
Forin Foundary & Mfg.
National Airoil Burner
Amsler Morton Div.
Bloom Engineering
York-Shipley Inc.
State
Location
Ohio
Pennsylvania
Remarks
Primarily boilers
Commercial sizes only
-
Primarily gas burners
Also made plant visit
No longer manufactuctures
Commercial sizes only
Boilers only
Commercial application
-
No longer makes oil burners
Commercial sizes only
Gas burners only
Commercial applications
Distribution for Wayne
Brochure
Received
No
No
Yes
Yes
Yes
No
No
No
Yes
Yes
No
No
No
Yes
Yes
243
-------
companies visited were members of NOFI and individuals within each company
were chairmen of various NOFI committees. The companies selected were
representative of major oil burner manufacturers in terms of sales and
research. However, these companies can all be classified as small busi-
nesses (<500 employees). This is understandable because the total market
for domestic oil burners is only about 800,000 units per year, of which
45 percent are new installations and 55 percent are replacement units.
The market should, however, increase in the future. In the past, appli-
cations have been limited primarily to heating units for homes, apartments,
and industrial and office buildings. New markets, which are growing each
year, are Marine applications and mobile home heating units. In addition,
because of the natural gas shortage, more and more existing units are
being converted to oil operation, and new installations are oil burners
alone. It appears that future heating units may be primarily oil burners,
rather than natural gas burners.
The companies selected for plant visits are tabulated in Table A-2. A
complete day was spent at each company in discussions with both management
and staff covering sales, production, engineering, and field service. The
areas of discussion were broad because the burner design is affected by
burner operation, cost, appearance, ease of service, as well as sales
promotions. The companies are largely controlled by sales considerations
and the function of engineering is to provide a product that is compatible
with the overall marketing strategy. For the most part, all of the oil
burners look alike and have identical locations for the pump, blower,
transformer, and controls. In fact, these component parts are frequently
identical on different burners, which facilitates obtaining replacement
parts anywhere in the country. The various burners are competitive cost-
wise and produce nearly identical efficiencies, with only a trace of smoke.
The major sales arguments are: (1) our product is more versatile and re-
quires less inventory (attractive to distributors or original equipment
manufacturers), (2) the product is tailored or engineered to your specific
application, (3) fixed geometry so that no adjustments are necessary, and
244
-------
Table A-2. LIST OF COMPANIES VISITED
K>
*.
cn
Company Name
The Carl in Co.
Sun Ray Burner Mfg. Co
Nu-lvay Division,
White- Rogers Corp.
R. h. Beckett Corp.
Automatic Burner Co.
Location
Ivethersfield, Conn.
Plainview, N. Y.
Milan, Illinois
Elyria, Ohio
Chicago, Illinois
Date of
Visit
2-12-73
2-13-73
2-14-73
2-15-73
2-16-75
Telephone
No.
(203) 529-2501
(516) 293-6800
(309) 787-4461
(216) 365-4141
(512) 735-170T
Initial
Contact
Mr. Burt S. Wat ling
Vice President
Mr. Bob Skarda
Sales Manager
Mr. Bud Rettke
General Manager
Mr. Myron Cooperrider
Chief Engineer
Mr. William Lord
Vice President
Co . Members
at Meeting
B. Watling-V.P.
Len Fisher-Chief
Engineer
Bob Skarda
(Chief Engr.
recently deceased)
Bob Krump-Supv.
Project Engr.
Les 01 sen Man-
ager Field
Services
M. Cooperrider-
Chief Engr.
R. Dumboys-Ass "t .
Chief Engr.
P. Double i Sales
0. Kaulhaber 'Mgr.
J. Beckett-Pres.
B. Cook-V.P.
h. Lord, V.I'.
-------
(4) factory-trained installation service to "fine tune" the burner. The
information and considerations relating to design practice, service prob-
lems, and independent research were almost the same for all companies, so
no attempt will be made to categroize by company.
The evolution of oil burner designs has been based on cut-and-try techni-
ques aimed at obtaining C02 levels of about 12.5 percent (with =20-per-
cent excess air) and no more than a trace of smoke. The oil burner de-
signs are now virtually fixed and little, if any, industrial manufacturer
research is being conducted to improve existing designs or develop new
approaches. Only two basic types of No. 2 oil burners are marketed today:
(1) flame-retention type and (2) conventional. The major difference be-
tween these types is that in the flame-retention type, a cone-shaped bluff
body or other obstruction is added that locally reduces the air and oil
velocities sufficiently below the turbulent flame speed so that ignition
is maintained at the retention device location. With the conventional
burner type, the flame location is dependent on burner and boiler design,
as well as operating conditions. Both types employ swirl vanes. In a
flame-retention-type burner, the swirl vanes are sometimes located on the
flame cone while, in the conventional burner, they are positioned along
the blast tube wall. An advantage of a centrally located flame cone with
an annulus between the cone and the wall is that it is possible to vary
the input air between the central portion of the flow area and the peri-
pheral zone and thereby adjust the mass distribution uniformity.
STATIC PLATE
Static plates are placed upstream of the oil nozzle and are designed to
reduce the air flow bias introduced by the centrifugal blower. The static
plates are primarily used to control air flow distribution when short
blast tube lengths are required, but with a pressure loss.
246
-------
BLAST TUBE
The length of the blast tube should be sufficiently long, usually in ex-
cess of 3 inches, to ensure that the air flow is uniform at the burner
head. The maximum length of blast tube currently used is about 20 inches.
The actual length is usually determined by the furnace wall thickness.
SWIRL VANES
The swirl vanes are used to promote mixing of the oil and air.
BLAST TUBE END PLATE
End plates (choke plates or choke rings) are used to induce a high com-
bustion gas velocity and thereby promote oil droplet breakup and addi-
tional mixing.
OIL NOZZLES
Primarily hollow cone nozzles are used in the smaller 0.6- to 3-gph burners.
These nozzles are designed to provide a very thin annular sheet of oil
with tangential swirl. The thinner the sheets, the smaller the final
droplets. Swirl promotes breakup and mixing with the air. The larger
burners use either hollow or solid cone, depending on specific character-
istics of each burner. At the low flowrate capacities, only a single
nozzle is used; however, for sizes larger than 12 gph, multiple nozzles
are sometimes used to control oil distribution and reduce dropsizes. The
multiple nozzle also is conducive to a low-fire, high-fire programmed
start sequence commonly found in the larger burners through use of one
nozzle for low-fire startup.
FLAME DETECTORS
Three types of flame detectors are currently being used as safety cutoff
devices; all are photoelectric devices. Cadmium sulfide cells are often
247
-------
used and ultraviolet-sensitive cells have been used. The devices operate
satisfactorily although ultraviolet-sensitive cell creates a mounting lo
cation problem, because it responds to ultraviolet-sensitive radiation
from the spark igniter and must be shielded from it.
SAFETY CONTROLS
If flame is not detected after the burner is turned on, the safety con-
trols must shut down the burner within an acceptable time. Underwriter's
Laboratories has specified the following times:
Capacity, Required Shutoff Time,
gph seconds
0.75 to 3.0 30
3.0 to 7.0 15
7.0 to 20.0 4
OTHER STANDARD PARTS
• Ignition rods--all standard; however, U.L. requires that the
electrodes be at least 1-1/2 inches away from any other metal
part
• Oil pump—100- or 300-psig discharge pressure
• Motor--1725- or 3450-rpm units
• Ignition Spark Transformer--10,000 to 12,000 volts
DESIGN PROBLEMS
The major problems that face the design engineer in burner/boiler matching
are:
1. Instability
a. Instability of the flame due to burner/boiler "mismatching"
or combustion coupling
248
-------
b. Instability induced from feed system vibrations
2. Noise
a. Pump or fan mountings and bearing vibrations which couple
easily with the boiler and associated ducts
b. Flame front and burner/boiler instabilities
3. Determination of insertion length (thickness of boiler wall and
the burner flange)
• Defines the blast tube length
4. Operation
• The burner must provide CO2 levels of 12.5 percent (with
20-percent excess air) with no more than a trace of
smoke at all design flowrates and, at its maximum rated
oil flowrate, it must demonstrate a 50-percent excess air
capability to allow for degradation.
Unfortunately, there is little, if any, interchange of design requirements
between the boiler and the oil burner manufacturers, so that the burner
designer is often faced with a combustion chamber configuration that will
not support efficient combustion.
The current practice in original equipment manufacturers' burner sales is
for the boiler manufacturer to ship a prototype boiler to the selected
oil burner manufacturer for use in testing at his plant. The oil burner
designer first inspects the boiler for furnace dimensions, refractory-
lined and uninsulated surfaces of the furnace, and required blast tube
length. With this information, a burner configuration is selected and it
is tested in the boiler furnace. Adjustments are made until the burner
operates satisfactorily. With some boilers, it has not been possible to
produce stable, efficient, noise-free operation, and the boiler design
has been rejected by both the burner and boiler manufacturer. In cases
249
-------
where instability is encountered, current design practice is to do one
of the following:
1. Change stack pressure to provide less flow resistance
2. Leave the boiler furnace door slightly ajar (boilers operate at
subatmospheric pressure)
3. Build in leaks around the furnace and burner installation flange,
which has the same effect as (2) above.
Although these fixes are crude, they work. Unfortunately, if the burner is
adjusted for efficient operation with no leaks, then introducing air leaks
will result in the boiler operating at a higher than desirable stoichio-
metric ratio. This condition usually does not result in increased smoke,
but the efficiency may be decreased.
A major problem facing the burner manufacturer is good field service. Gen-
erally, the manufacturer provides no additional service after the burner
leaves the plant. Usually, the boiler manufacturer or distributor provides
service installers whose job is to ensure that the burner is properly set
up. The burner people universally agreed that service and installation
are generally poor. Service personnel often have only a minimal knowledge
of good operating practice, and no standards are maintained. The burner
manufacturers indicated that the units should be removed and serviced at
least once a year to prevent fouling of the nozzle, carbon deposits on
the burner, and ensure that the igniters are working properly. Owners,
however, generally service their burners only when they fail. The oil
burner manufacturers are not inclined to change their burner designs be-
cause they believe that field service personnel are reluctant to learn new
procedures and stock additional replacement parts. This is the reason why
new innovations should be relatively straightforward modifications of
existing designs.
250
-------
APPENDIX B
FACTORS FOR THE CONVERSION OF UNITS TO THE METRIC SYSTEM
Physical
Quantity
Length
Mass
Pressure
Speed
Temperature
Time
Volume
To Convert
From
inch
foot
micron
pound
psi
newton/meter2
foot/second
Celsius (C)
Fahrenheit (F)
Fahrenheit (F)
Rankine (R)
hour
gallon (U.S. liquid)
To
meter
meter
meter
kilogram
pascal
pascal
Multiply
by
Kelvin (K)
Kelvin (K)
Celsius (C)
Kelvin (K)
second
meter3
0.0254
0.3048
0.000001
0.45359
6894.8
1.0000
meter/second 0.3048
K = C + 273
K = (F + 460)/1.8
C = (F - 32)/1.8
K = R/1.8
3600
0.0037854
251/252
-------
TECHNICAL REPORT DATA
(PUmc read luan'u luun on ilic rci i rsi IH lure romplrluifl
1. RLi'OiH NO.
EPA-650/2-74-047
3. RtUPILNT'S ACCtbZilOftNO.
> TITLL AND SUBTITLE
Design of an Optimum Distillate Oil Burner for
Control of Pollutant Emissions
S RCPOFH DATE
June 1974
G. PERFORMING ORGANIZATION CODE
.AUTMOR(S)
. A. Dickerson and A. S. Okuda
B PERFORMING ORGANIZATION REPORT NO
PERFORMING ORGANIZATION NAME AND ADDRESS
^ocketdyne Division,
Rockwell International
5633 Canoga Avenue, Canoga Park, CA 91304
10. PROGRAM ELEMENT NO
1AB014; ROAP 21ADG-44
11. CONTRACT/GRANT NO
68-02-0017
12. SPONSORING AGENCY NAME AND ADDRESS
SPA, Office of Research and Development
1ERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TVPC OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
S. SUPPLEMENTARY NOTES
16 ABSTRACTrpne rep0rt describes results of a research study of the pollution character-
istics of high-pressure atomizing, No. 2 distillate fuel oil burners. The main empha-
sis was on optimizing burner design to minimize pollutant emissions when firing into
refractory-lined combustion chambers. The atomizing characteristics, and flow and
composition profiles in the combustion zones of several commercial burners were
determined experimentally. Mass median droplet diameters were 60-90 microns for
. 50-1. 50 gph oil spray nozzles. Nitric oxide (NO) formation was most prevalent in
he near-stoichiometric combustion zones where local flow conditions led to vigorous
as mixing. These data were used to design variable geometry burners, used to
timize burner geometry for minimizing pollutant emissions. The optimum geometry
lurners were fabricated in fixed-geometry versions and tested extensively to verify
their low air pollutant emissions. Substantial reductions (about 50 percent) in NO
emissions were achieved by optimizing conventional designs , with negligible emis-
sions of other pollutants. Also, several nonconventional burner designs were built and
tested: two of these led to very low NO emissions. Program results have been used
to develop recommendations for burner design to minimize pollutant emissions.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b IDENTIFIERS/OPEN ENDED TERMS C. COSATI I wld/ClOUp
Air Pollution
Combustion
Oil Burners
Design
Nitrogen Oxides
Fuel Oil
Distillates
Heating Equipment
Air Pollution Control
Stationary Sources
Burner Design
Oil Spray Characteri-
zation
Flow Field Measurement
13B
21B
13A
07B
11H, 21D
18. DISIRIUUTION STATEMENT
Unlimited
10 SECURITY CLASS I Tins Report)
Unclassified
21 NO Of PAGES
270
JO SECURITY CLASS
Unclassified
EPA Form 2220 I (9-»l|
253/254
------- |