oERA
United States Industrial Environmental Research EPA-600/7-80-043
Environmental Protection Laboratory March 1980
Agency Research Triangle Park NC 27711
Pilot Scale Combustion
Evaluation of Waste and
Alternate Fuels:
Phase III Final Report
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-043
March 1980
Pilot Scale Combustion Evaluation
of Waste and Alternate Fuels:
Phase III Final Report
by
R.A. Brown and C.F. Busch
Acurex Corporation
Energy and Environmental Division
485 Clyde Avenue
Mountain View, California 94042
Contract No. 68-02-1885
Program Element No. EHE624A
EPA Project Officer: David G. Lachapelle
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
This report gives results of three studies at EPA's Multifuel
Test Facility. The first evaluated a distributed-air staging concept
for NOV control in pulverized-coal-fired systems. The results showed
A
that minimum NO levels of 140 ppm were achieved at overall residence
times similar to those used during conventional staging tests. However,
the NO levels achieved with the distributed-air concept were no lower
than those achievable with conventional staging. The second evaluated
combustion control techniques and NO emissions when firing coal/oil
mixtures. NO emissions for a given burner and nozzle were generally
proportional to the fuel-nitrogen content of the fuel. Additionally,
combustion control technology currently used for NO control from pul-
A
verized coal was found to be effective with coal/oil mixtures, but to
differing degrees, depending on the coal/oil mixture ratios and compo-
sitions. The third evaluated emissions and combustion characteristics
of refuse-derived fuel (RDF) co-fired with either natural gas or pul-
verized coal. Four RDF materials were evaluated for gaseous, particulate,
trace metal, and organic emissions. In general: CO and UHC emissions
were low; NOX and SOX emissions decreased with increasing RDF content
when co-fired with coal; particulate levels did not substantially in-
crease with the RDF; and no trace metal emissions correlation was found.
iii
-------
CONTENTS
Abstract in
Figures vi
Tables vii
Conversion Table xii
1. Overview and Summary 1
2. Distributed Air Tests 7
2.1 Special experimental hardware 9
2.2 Test plan 15
2.3 Experimental data 18
2.4 Conclusions 35
3. Coal/Oil Mixture Tests 37
3.1 Test plan 50
3.2 Test data 54
3.3 Baseline tests 56
3.4 Control technology tests 60
3.5 Fuel nitrogen studies 69
3.6 Burner nozzle comparison 71
3.7 Previous testing data 71
3.8 Summary and conclusions 71
3.9 Recommendation 75
4. Refuse-Derived Fuel Tests 76
4.1 Objectives 77
4.2 RDF experimental hardware 78
4.3 Test theory and plan . . 101
4.4 Analytical procedures 106
4.5 Experimental data 114
References 166
Appendices
A. Data summary - distributed air; coal/oil mixture; RDF 169
testing; COM/DOE report
-------
FIGURES
Number Page
Distributed Air Tests
2-1 Distributed air concept as per Pershing 8
2-2 Distributed air arrangement in the horizontal extension ... 10
2-3 IFRF burner 12
2-4 B&W-type coal spreader 13
2-5 Baffle detail 14
2-6 Residence time in section Ib 20
2-7 Distribured air configurations 21
2-8 Effect of SR]a 24
2-9 Effect of SRla 25
2-10 Effect of SR]b 27
2-11 Effect of la stage residence time 28
2-12 Effect of la stage residence time 29
2-13 Effect of Ib stage residence time 31
2-14 Effect of Ib stage residence time 32
Coal/Oil Mixture Tests
3-1 EPA/Acurex multifuel furnace 39
3-2 Facility modifications 40
3-3 Delavan swirl-air nozzle 41
3-4 Sonic Corporation Sonicore nozzle 42
VI
-------
FIGURES (CONTINUED)
Number Page
3-5 Coal/oil delivery system 49
3-6 Baseline emissions 58
3-7 Boiling point curves 59
3-8 Nitrogen evolution curves 61
3-9 Staging emissions 62
3-10 Burner air distribution 64
3-11 Burner air distribution 65
3-12 Air distribution tests 66
3-13 Effect of firing rate and residence time 68
3-14 Effect of fuel nitrogen 70
3-15 Nozzle comparison 72
3-16 Date from earlier work 73
Refuse-Derived Fuel Tests
4-1 Furnace cross section 79
4-2 Tangential configuration, aerodynamic patterns 80
4-3 Corner-fired burner 81
4-4 RDF nozzle 82
4-5 Modified corner-burner assembly 84
4-6 Fuel delivery schematic 85
4-7 RDF feed system design 87
4-8 Pneumatic transport system 88
4-9 Safety system 90
4-10 Sampling system online at experimental multiburner furnace . 95
4-11 Aerotherm high volume stack sampler 96
vii
-------
FIGURES (CONTINUED)
Number Page
4-12 Source assessment sampling system (SASS) 97
4-13 Ash deposition 102
4-14 Test matrix for baseline emissions characterization .... 104
4-15 Test matrix for emissions control through theoretical
air variation 105
4-16 Test matrix for baseline emissions characterization .... 107
4-17 Test matrix for emissions control through theoretical
air variation 108
4-18 Photographs of fuel samples 117
4-19 NO emissions during baseline testing (Ames) 120
4-20 NO emissions during baseline testing (Richmond) 121
4-21 NO emissions during baseline testing (Americology) 122
4-22 NO emissions during baseline testing (San Diego) 123
4-23 Thermal NO (previous work) 124
4-24 NO emissions during baseline testing (all RDF's) 126
4-25 S02 data (all RDF's) 127
4-26 NO emissions during detailed testing (Richmond RDF/
Pittsburgh coal 129
4-27 Fuel nitrogen contribution 130
4-28 Stack gas particle size vs. cumulative percent less
than diameter 137
4-29 Particulate loading results 138
4-30 Stack gas particulate size vs. cumulative percent
less than diameter - trace metal Cu 149
4-31 Stack gas particle size vs. cumulative percent
less than diameter trace metal Zn 150
vm
-------
FIGURES (CONCLUDED)
Number Page
4-32 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Mn 151
4-33 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Pb 152
4-34 Stack gas particle size vs. cumulative percent
less than diameter trace metal Cd 153
4-35 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Be 154
4-36 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Ti 155
4-37 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Sb 156
4-38 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Sn 157
4-39 Stack gas particle size vs. cumulative percent
less than diameter - trace metal Hg 158
4-40 Stack gas particle size vs. cumulative percent
less than diameter - trace metal As 159
IX
-------
TABLES
Number Page
Distributed Air Tests
2-1 Optimum Staging Parameters (Pershing) 9
2-2 Range of Parameters of Interest 16
2-3 Revised Distributed Air Studies Matrix 17
2-4 Residence Times in la Stage 19
2-5 Distributed Air Versus Conventional Staging 34
2-6 Facility Characteristics 35
Coal/Oil Mixture Tests
3-1 Emission Monitoring Equipment 44
3-2 Fuel Oil Analyses 45
3-3 Coal Analyses, As-Received Basis 46
3-4 Coal/Oil Mixture Analyses, As-Received Basis, 30% Coal
by Weight 47
3-5 Baseline Matrix 51
3-6 Effect of Load and Residence Time 52
3-7 Distributed Air Burner Tests 53
3-8^ Effect of Fuel Nitrogen 55
Refuse-Derived Fuel Tests
4-1 Emission Monitoring Equipment 94
4-2 Fuel Analysis 106
-------
TABLES (CONCLUDED)
Number Page
4-3 Metals Which Were Analyzed ................. 109
4-4 Liquid Chroma tog raphy Elution Sequence ........... 110
4-5 Distribution of Compound Classes in Liquid Chroma tographic
Fractions of Organic Extracts .............. Ill
4-6 Test Matrix ......................... 115
4-7 Fuel Analyses ....................... 116
4-8 Parti cul ate Analyses: Effect of RDF Type ......... 131
4-9 Particulate Analyses: Effect of Excess Air ........ 133
4-10 Particulate Analyses: Effect of Percent RDF ....... 135
4-11 Particulate Analyses: Coal vs. 10% RDF + Coal vs.
RDF + Gas ....................... 136
4-13 Total Trace Metal Loadings -Coal Cofiring ......... 142
4-14 Trace Metal Concentrations as Vapor - Coal Cofiring .... 144
4-15 Total Trace Metal Loadings -Gas Cofiring ......... 145
4-16 Trace Metal Concentrations as Vapor - Coal Cofiring .... 146
4-17 Trace Metal Concentrations - Pilot vs. Full Scale -
Particulate Only ..................... 148
4-18 Organics Found ....................... 161
4-19 LC Column Data ....................... 163
4-20 LC Column Data ....................... 164
4-21 Possible Compounds in LC Fractions not Analyzed ...... 165
XI
-------
CONVERSION TABLE
ENGLISH TO SI METRIC CONVERSION FACTORS
To convert from To Multiply by
inch m 2.540 000 E-02
foot m 3.048 000 E-01
scfm m3/s 4.719 474 E-04
gr/ft3 g/m3 2.288 352 E+00
gallon m3 3.785 412 E-03
gph m3/s 1.051 503 E-06
Btu/hr Wt 2.930 711 E-01
Ib/hr kg/s 1.259 979 E-04
Btu/lb kJ/kg 2.326 000 E+00
Btu/hr-ft3 Wt/m3 1.034 971 E+01
Ib kg 4.535 924 E-01
psig kPa 6.894 757 E+03
ug/Btu yg/J 9.478 170 E-01
°F °C t°C = (t°F-32)/1.8
scfm = standard cubic feet per hour
gr/ft = grains per cubic foot
gph = gallons per hour
Btu = British thermal unit
hr = hour
Ib = pound
ft3 = cubic foot
psig = pounds per square inch (gauge)
yg = micrograms
m = metre
3
m = cubic meter
s = second
g/m = grams per cubic meter
Wfc = thermal watt
kg = kilogram
kJ = kilojoule
kPa = kilopascal
xii
-------
SECTION 1
OVERVIEW AND SUMMARY
The work summarized in this report was performed during the period
October 1977 to July 1978 as Phase III of the Pilot Scale Evaluation of
N0₯ Combustion Control Techniques, EPA Contract 68-02-1885. This report
A
discusses
Advanced NOY Control Techniques for Pulverized Coal Through
A
Distributed Air
Emissions and NO Control Technology Evaluation of Coal/Oil
rt
Mixtures (COMs)
Evaluation of Emissions on Co-firing of Four Refuse-Derived
Fuels (RDF) with Natural Gas and Pulverized Coal
A brief summary of the scope and results from each of these tasks follows.
Distributed Air Tests
Tests at the University of Arizona in a bench-scale coal-fired fur-
nace suggested that low NO levels could be achieved in relatively overall
short residence time by sequencing the air into the burner in three stages.
The primary zones include the primary air conveying coal and some secondary
air. The stoichiometric ratio of this first stage would be in the range
of 0.3 to 0.6 and a residence time of 0.3 to 0.75 seconds. Tertiary air
It is the EPA policy to use SI metric units; however, in this report
English units are occasionally used for convenience. See attached con-
version table.
-------
was then added through four ports at 90° to the flow of combustion products.
This second stage is held at a stoichiometric ratio of 0.75 to 0.95, and
staging air is then added through four additional ports.
This series of tests explored a range of stoichiometric ratios and
residence times for each stage. Unfortunately, the results from these
tests could not duplicate the Arizona tests. What was found was that even
with the distributed air approach, NO levels increase with decreasing SR
below an SR of about 0.6. In addition, NO always decayed with increasing
residence time in both the first and second stages. Minimum NO levels of
about 140 ppm were achieved in overall residence times similar to the con-
ventional staging results. Thus, no advantage over conventional staging
was achieved. These results may be partially explained by the fact that a
diffusion burner was used in a relatively low L/D firebox as compared to a
premixed burner in a high L/D firebox in the Arizona tests.
Coal/Oil Mixture Tests
In order to utilize coal in the near term, it has been proposed to
fire oil or gas boilers with a slurried mixture of coal and oil. Although
the feasibility has been demonstrated in a number of small and larger scale
demonstrations, there is a need to determine the environmental problems
associated with COMs. Because of the generally higher fuel-N content of
the mixture as compared to oil, it seems likely that the NO levels would
/\
also be higher. Therefore, the purpose of this study was as follows:
Obtain emission data for coal/oil combustion in an environment
closely simulating an industrial package boiler
Determine if emissions levels were affected by the fuel com-
position
Determine if conventional control technology developed for coal
-------
is effective in reducing emissions levels produced by coal/
oil combustion
Investigate the effect of burner modification on emission
levels produced by coal/oil combustion.
During this study, two oils and three coals were fired in a package
boiler simulator. Baseline emissions of the parent fuels and the fuel com-
binations were determined. Control technology tests were run on the COMs
as follows:
Baseline NO emissions from COM were, in general, proportional
to the fuel-N for a given burner and nozzle type.
t The burner settings and fuel nozzle type have a strong effect
on NO emissions.
Conventional control technology currently utilized for pulver-
ized fuel combustion is effective in reducing NO emissions, but
to different degrees, depending on fuel composition
NO emissions increase in proportion to the amount of coal in
the coal-oil mixture, but fall between the parent fuel oil and
coal baseline emissions.
If there is to be significant utilization of COM in industry, and if NO
levels are to be controlled, much additional work is needed to understand
the mechanisms which control NOV formation in COMs for different fuel
A
combinations.
Refuse-Derived Fuels Testing
It is necessary that investigations regarding the environmental
compatibility of RDF be conducted before this vast, untapped energy source
can be considered a viable supplement to present energy resources. It has
thus been suggested that such investigations can be carried out most cost
-------
effectively in a pilot-scale facility. To determine the feasibility of
such pilot-scale testing, the IERL of EPA/Cincinnati funded a study as
part of the Phase III activity. The goals of this study were as follows:
To design, fabricate, and operate a system for combustion
testing of RDF in a laboratory scale facility
0 To characterize RDF emissions of several types of material
presently available for use as fuel
To evaluate the combustion efficiency of fuels consisting of
conventional clean and dirty fossil fuels (natural gas and coal)
mixed with varying percentages of refuse
To evaluate the effects of combustion parameters on emissions
from RDF/conventional fuel mixtures
t To provide direction for future investigations on refuse-1 derived
fuel to insure solutions to problems associated with its use.
A feed system was designed to control and measure from 10 to 60
Ib/hr of a variety of "fluff" refuse-derived materials. This was accom-
plished using a rotating drum hopper depositing on an internal moving belt
conveyor. The conveyor, in turn, deposits the material into a tube where
it is conveyed by air into the top port of a tangential burner. The pilot-
scale facility was tangentially-fired at 1.5 x 106 Btu/hr with RDF fed to
two of the four corners. The test program was designed to determine the
gaseous, particulate, trace metal, and organic emissions of four sources
of RDF. The four materials were from San Diego, California; Richmond,
California; the Americology Facility in Milwaukee, Wisconsin; and from the
Ames, Iowa Plant. NO emissions increased with both percent RDF and percent
excess air when fired with natural gas. NO emissions also varied for the
-------
four fuel types in approximate proportion to the fuel-N content of the RDF.
When co-fired with coal, the NO emissions decreased with increasing percent
RDF, even though the percent N available increased. This may be attributed
to shielding of the coal by the RDF from the oxygen, flame/flame processing
or locally fuel-rich zones in the coal stream caused by redistribution of
the combustion air during the RDF firing.
Particulate levels did not substantially increase with the RDF, but
a higher concentration was found in the less than ly size range. No corre-
lations were found with trace metal emissions, either with respect to
percent RDF or percent excess air. It is believed that, due to the great
variability in the feed from minute to minute and the problems with holdup
in the heat exchanger sections, a valid trace metal evaluation is not pos-
sible from a single test. Many tests will be necessary to form a statisti-
cally reliable number. Lastly, few poly-organic materials (ROMs) were found
and no poly-chlorinated biphenyls (PCBs).
Combustion efficiency of these pilot-scale tests was perhaps better
than full-scale tests due to higher combustion temperatures. Additional
tests are needed to better determine the variability of gaseous, trace
metal, and organic emissions from a single source and from a variety of
sources.
In the sections that follow, the results from each of these test
programs will be explained in more detail. In all cases they were per-
formed in EPA's multifuel, multiburner test facility located at Acurex
Corporation in Mountain View, California. The facility was either used
in its normal utility boiler configuration (main firebox only) or with
the horizontal extensions which can simulate a package boiler configuration
-------
or serve as an asymmetric flow combustion system. Details of this facility
may be found in the Phase II report (Reference 1). A description of the
special equipment required for each of the test programs and details of
the configurations for those tests are included within the section on
that program.
-------
SECTION 2
DISTRIBUTED AIR TESTS
The conventional staging studies performed in the main firebox and
horizontal extensions (Reference 1) of the EPA Multiburner Test Facility
revealed that low NO levels could be achieved with a sufficiently long
A
residence time under fuel-rich conditions. However, this approach neces-
sitates an exceptionally long residence time under fuel-rich conditions
with potential for corrosion and slagging problems. Thus, an alternate
approach was sought.
Tests run on a subscale premixed combustor at the University of
Arizona (Reference 2) indicated that a three-stage approach would be able
to achieve low NOX levels in an overall residence time of less than 1.5
seconds.
The approach is illustrated for the Arizona facility in Figure 2-1.
In this arrangement premixed coal, primary air, and secondary air enter
the top of the furnace at a stoichiometric ratio SR-|a. This stoichiometry
is held for a residence time T.J seconds, whereuopon tertiary air is intro-
duced. The second part of the first stage is held at a stoichiometric
ratio of SR^b for i^ seconds. The second-stage air is then introduced
for final burnout under excess air conditions for a residence time of T2
seconds. Pershing found that for his facility a unique combination of
-------
1st
Stage
2nd
Stage
Coal + primary air
+ secondary air
la
Tertiary
air
la
lb
Second
stage j-
air
T2
Flue
Figure 2-1. Distributed air concept as per Pershing,
-------
SR, , T, , SR,. , and T,. achieved the lowest NOV level. The optimum condi-
ia ia iD ID x
tions for the Pershing experiment are tabulated in Table 2-1. In summary
he found the following:
Stack NO decreased with decreasing SR, until an SR-|a = 0.5.
Below an SR, = 0.5, stack NO did not decrease further for con-
stant conditions downstream
An optimum T, was necessary to achieve minimum NO . On either
i a x
side of the optimum, NO would increase.
/\
NOX decreased with decreasing SR,b
NOX increased with decreasing T-,.
» T? residence time had no effect on the stack NO
L- A
TABLE 2-1. OPTIMUM STAGING PARAMETERS (PERSHING)
SRla
SRlb
Tla
Tlb
0.4 - 0.5
0.85
0.4 sec
1-1.5 sec
(not varied)
This approach was thus investigated using the EPA multiburner-horizontal
extension test facility.
2.1 SPECIAL EXPERIMENTAL HARDWARE
It was decided that the horizontal extensions would be the best
equipment to perform these tests. The horizontal extensions were set up
as shown in Figure 2-2. Each horizontal extension is 33-inch inside diam-
eter, refractory lined, and two feet long. Up to five sections may be
-------
TERTIARY AIR
STAGED AIR
SECONDARY AIR
PRIMARY
+ FUEL
PRIMARY
+ FUEL
SECONDARY AIR
SRioT - 1.15
Figure 2-2. Distributed air arrangement in the horizontal extension.
-------
joined together to form an overall length of 10 feet. A transition section
connects to the main firebox where the flue gases are then quenched by the
.~_r heat exchange sections. Gaseous emissions are sam-
pled just downstream of the heat exchange section. On the other end of the
horizontal extensions, either a single burner or up to five burners may be
mounted. For this test, four of the nominal 300,000 Btu/hr IFRF variable
swirl block burners were chosen. These burners were fitted with a 2-inch
diameter sleeve in the air throat to achieve reasonable velocities under
very fuel-rich conditions. The burner is illustrated in Figure 2-3. The
burners also used the B&W-type coal spreader illustrated in Figure 2-4,
and the swirl was set at a mid-position of four. Part of the objective
for utilizing these four burners in this configuration was to achieve a
well-mixed first stage. To further enhance the mixing in each stage, baf-
fles were used whenever possible at the end of the la stage and Ib stage,
as was illustrated in Figure 2-3. These baffles also served to separate
the stages and prevent backmixing into the first stage. The baffle or choke
was made from a high temperature refractory in four sections as shown in
Figure 2-5. This arrangement made it relatively easy to move the baffle to
any desired location within the tunnel. At the first tertiary air position,
it was not possible to install the baffle due to the close proximity to
the burners. The baffle opening was 16 inches in diameter. The tertiary
air was introduced just downstream of the first baffle in four locations
90 degrees apart. This air was introduced through 2-inch diameter ports
perpendicular to the main flow. The horizontal extensions have four ports
90 degrees apart every foot along the length of the furnace. The first
four locations, 1 foot apart, were chosen to vary the tertiary air residence
11
-------
ro
Figure 2-4. IFRF burner.
-------
27.5°,
Figure 2-4. B&W-type coal spreader.
-------
Figure 2-5. Baffle detail.
-------
time t, . This enabled the la stage residence time to be varied from
i a
less than 0.5 second to over 2 seconds. The rlb residence time was then
varied by positioning the staging air ports from 2 to 7 feet downstream of
the tertiary air position. No attempt was made to control the second-stage
residence time, but it was always sufficient to complete combustion.
Temperature of the secondary and tertiary air was maintained at about
600°F and the stage air at 300°F. Bare platinum-platinum/rhodium thermo-
couple measurements were made in the la stage and Ib stage.
2.2 TEST PLAN
The tests were structured to explore the following variables in the
distributed air concept:
la stage residence time T,
Ib stage residence time T-,.
la stage stoichiometric ratio SR-,
0 Ib stage stoichiometric ratio SR,.
Firing rate
Temperature
The range of each parameter of interest is given in Table 2-2.
Table 2-3 shows the matrix that was run. This matrix is the matrix which
was developed ;during the course of the testing as the results redirected the
effort. It was found, for instance, that a fairly dense matrix was needed
to truly see the effects of the various parameters. In general two to
three la stage residence times and two Ib residence times were selected
for each tertiary air position. The la stage stoichiometric ratio was
varied from 0.3 to 0.7, and two Ib stoichiometries of 0.85 and 0.95 were
selected. The firing rate changes not only effect early mixing, but the
15
-------
TABLE 2-2. RANGE OF PARAMETERS OF INTEREST
Parameter
Range
SR
la
SRlb
Tla
Tlb
Firing Rate
Temperature
0.3 -0.7
0.8 -0.95
0.5 - 3.0
0.5 - 1.5
0.85 - 1.7 x 106 Btu/hr
Ambient - 600°F
16
-------
TABLE 2-3. REVISED DISTRIBUTED AIR STUDIES MATRIX3
Distributed
Air Studies
Matrix
r*
r
S-
C.
~>.
3
4*
CO
PI
u> «
O .
a
nj
01
£
d.
U-
o
O
O
*l«S
s
£
Q.
LL.
§
+*«>
3
V
l-
o_
CD
5
in
tn
^j°
ae.
on oo
0
in
en
J3°
ee.
oo oo
d
in
CTl
^o
cc.
00 CO
o
in
en
^i°
cc.
00
00
0
Position #1 - T,
tlb - Short
0.3
209b
(b)
S^
0.45
209a
(b)
a
0.6
209c
(b)
T - Long
0.3
209J
a)
209g
209t
a)
SRla
0.45
2091
(a)
209k
(a)
209h
(a)
209e
(a)
0.6
2091,
(a)
209d
(a)
Position #2 - T]fl
T,. - Short
0.3
21 3x
(f)
21 3u
(f)
21 Si-
te)
21 3 j
(e)
213d
(e)
21 3c
e)
SRla
0.45
21 3v
(f)
213s
(f)
21 3h
(e)
213k
(e)
21 3e
(e)
21 3a
(e)
0.6
21 3w
(f)
21 3t
(f)
213g
(e)
213*
(e)
213f
(e)
21 3b
(e)
tlb - Long
0.3
21 3p
(f)
213o
(f)
SRla
0.45
21 3q
(f)
21 3n
(f)
0.6
213r
(f)
213m
(f)
Position #3 - T,
T,. - Short
ID
0.3
21 2J,
(k)
211b
(k)
21 2g
(h)
211j
(h)
212f
h)
212b
h)
SR1a
0.45
212k
(k)
211a
(k)
21 2h
(h)
211k
(h)
21 2e
(h)
21 2a
(h)
0.6
21 2j
(k)
211c
(k)
212i
(h)
21H
(h)
21 2d
(h)
212c
(h)
Tlb - Long
0.3
21 Ob
(j)
21 Of
(J)
211i
(k)
211f
(k)
SRla
0.45
21 Oa
(j)
210e
(j)
211g
(k)
211d
(k)
0.6
21 Oc
(J)
21 Od
(j)
211h
(k)
211e
(k)
Position *4 - T,
1
-------
residence time between any two points in the furnace and the temperature at
any point. A few tests were also run with no preheat and the lowest firing
rate to further enhance any temperature effect. When no effect was seen
at this lowest load, further tests of temperature were exchanged for a more
complete matrix in other regions.
To aid in estimating the effect of residence time, a plot and tables
were prepared to determine the residence time for a given staging position,
firing rate, and stoichiometric ratio. Table 2-4 presents this data for
the la residence time as a function of tertiary air position, stoichiomet-
ric ratio, and firing rate. Figure 2-6 presents residence time plots for
the Ib stage as a function of firing rate and SR^, . The absissa of this
plot is the distance between the tertiary air position and the stage air
position. Also shown on this plot are the various configuration letters
associated with each test point. The matrix in Table 2-3 gives the configu-
ration letter for each test number, and the configurations are depicted in
Figure 2-7. Thus, by referring to the matrix, the configuration letter can
be determined and seen schematically in Figure 2-7. By referring to Figure
2-6 and knowing the firing rate and SR,. , the residence time in the Ib stage
may be determined. This procedure together with the table for T, were used
to reduce the data to common residence times. Thus, the true effect of the
stoichiometric ratios and, conversely, the effect of residence time at con-
sistent stoichiometric ratio could be determined.
2.3 EXPERIMENTAL DATA
The data have been compiled in this section in its reduced form.
First, all data were reduced to an air-free (0 percent 02) basis. Then
cross plots were made so that the true effect of any parameter could be
18
-------
TABLE 2-4. RESIDENCE TIMES IN la STAGE
(T, seconds)
IT
-C
TJ 3
(0 -P
O CO
(A
0
1.7
1.3
0.85
Position 1
SRla
0.3
0.487
0.637
0.974
0.45
0.354
0.464
0.711
0.6
0.280
0.366
0.560
Position 2
SRla
0.3
0.885
1.16
1.77
0.45
0.644
0.84
1.29
0.6
0.509
0.67
1.02
Position 3
SRla
0.3
1.47
1.93
2.94
0.45
1.06
1.39
2.14
0.6
0.84
1.11
1.69
Position 4
SRla
0.6
1.17
--
0.7
1.10
--
-------
ro
O
2.8
2.6
2.4
2.2
2.0
1.8
o> 1.6
E
a, 1.4
c
0) ,
-a 1.2
in
-------
External configuration
Internal configuration
"b"
"d"
"h"
i
}
* J
1
1
1
i
i
\
i
i
i
i
i
J *
i
i
i
i
i
i
i
i
i
. t
i
't *
*
*
i
Y t *
i
i
1
1
1
1
Figure 2-7. Distributed air configurations,
21
-------
Configuration
number
"n"
1
1
1
t
1
1
1
t
1
i
1
1
t
1
1
1
t
*
1
1
t
1 1
1
1. 1
1
I 1
~~*
T
o
o
o
1 i
\ 1
o
o
0
Figure 2-7. Distributed air configurations (concluded)
22
-------
determined. Each of the various parameters will be discussed in the sub-
sections that follow.
2.3.1 la Stage Stoichiometric Ratio
Figure 2-8 shows the effect of the la stage Stoichiometric ratio at
a constant SR^ of 0.80, la stage residence time of 1.00 seconds, and t.,
residence time of 0.97 seconds. As can be seen on the plot, the effect was
fairly pronounced with NO decreasing with increasing SRla. Previous tests
indicated that above an SR of 0.7-0.8, the NO levels would again increase.
This result is in contrast to the Pershing result which indicated that there
was no effect of SR. below a level of SR, 0.50. Actually, this data is
consistent with the data taken previously in this facility with conventional
staging. It was suggested in that previous work that this increase in NO at
low Stoichiometric ratios was due to the formation of .second-stage NO.
A similar trend was found at an SR,. = 0.95 as seen in Figure 2-9.
There is some question as to the validity of the points at 1.7 x 106 Btu/hr,
particularly at the lower SR, 's because the baffle collapsed during these
tests. Nevertheless, the data still indicate that NO increases with de-
creasing Stoichiometric ratio. The NO levels are about 100 ppm higher,
however, at an SRlb of 0.95 than at an SRlb of 0.80. This suggests that
some second-stage NO is being formed and is either dependent on the Stoi-
chiometric ratio in the Ib stage or on a dependent variable of the Stoi-
chiometric ratio. Recent tests (Reference 3), for example, have shown that
second-stage NO is strongly dependent on the initial flame temperature in
the second stage, provided this temperature is below 2200°F. However, if
it was an effect of temperature, in this case we would expect to see an
increase of NO with firing tate. In both plots this is true at an SR
23
-------
600
500
400
C\J
0 300
o
o
200
100 ~
0.30
T, = 1.0 sec
I a
»lb ^ 0.97 sec
R - 0.80
Load
O 1-7 x 10 Btu/hr
A 1 .3 x 10 Btu/hr
D 0.85 x 10 Btu/hr
_L
0.40 0.50
SRla
0.60
0.70
Figure 2-8. Effect of
-------
ro
en
600
500
400
300
200
100
Effect of SR]a
T.J = 1.0 sec
tlb = 0.97 sec
- 0.95 sec
O 1 .7 x 10" Btu/hr
A 1 -3 x 10*' Btu/hr
D 0.85 x 10f Btu/hr
0.30
0.40
0.50
0.60
0.70
SR
la
Figure 2-9. Effect of SRlg(.
-------
of 0.3, but not at 0.6. Previous data (Reference 1) indicated the higher
the first-stage temperature, the greater the decay rate in the first stage.
Thus, it is possible that there are competing effects with an optimum decay
at the high load and 0.6 stoichiometric ratio, while at the 0.3 SR-, greater
heat release is experienced in the second stage causing the second-stage NO
to increase. In fact, this is a possible explanation for the decrease of
NO with increasing stoichiometric ratio up to an SR, of 0.7. (It is known
i a
from previous work that the NO will again increase beyond an SR of 0.7).
2.3.2 1b Stage Stoichiometric Ratio
The effect of SRlb is shown in Figure 2-10 for an SRla of 0.3 and 0.6.
As can be seen, NO always increased with an increase in SR,, . Again, this
may be either due to the greater degree of oxygen availability or may re-
flect an increase in flame temperature in the Ib stage. This data is also
interpolated data at constant residence times in the la and Ib stages.
2.3.3 Residence Time, la stage
The effect of the la stage residence time is shown in Figure 2-11
and 2-12 for an SRlb of 0.80 and 0.95, respectively, and an SRlg of 0.6
(the point at which minimum NO occurred). The data for the three firing
rates is also included on these plots. These data are stack emisions at
an overall excess air level of 15 percent. The data indicates a strong
decay between 0.5 to 1.0 second at all loads and then a decrease in the
decay rate following 1.0 second. The data also suggest that at the higher
load the initial NO levels are higher, but the decay rate is also higher
resulting in lower NO levels after a T, of 1.0 second. This is consistent
with previous results which show lower NO levels at higher load or first-
stage temperatures. The main conclusion from this curve is that little
26
-------
ro
C\J
o
o
o
600
500
400
300
200
100
= 1.0 sec
= 0.97 sec
SRla = 0.3
SRn
O 1-7 x 106 Btu/hr
A 1.3 x 106 Btu/hr
D 0.85 x 10* Btu/hr
0.80
0.95
SR
Ib
Figure 2-10. Effect of SR,
-------
PO
oo
600
500
400
Q.
Q_
C\J
O
300
200
TOO
Western Kentucky coal
4 IFRF burners
Smallest sleeve
B&W spreader, SW = 4
Excess air = 15;
SR] = 0.60
SRit, = 0.30
= 0.97 sec (all values converted
Tlb
to this common T,, )
O 1.7 x 10" Btu/hr
Q 1.3 x 106 Btu/hr
0 0.85 x 106 Btu/hr
1 1
1
1
1 1 1 1
1 1
0.4 0.8 1.2 1.6 2.0
Tla (56C)
2.4
2.8
3.2
3.6
Figure 2-11. Effect of la stage residence time.
-------
600
500
400
ro
vo
Q.
CL
o
o
200
100
0.4
Western Kentucky coal
4 IFRF burners
Smal lest sleeve
B&W spreader, SW = 4
Excess air = 15,*
= 0.60
- 0.95
= 0.97 (all
values converted
to thi s
common i ., )
I
0.8
1.2
1.6
2.0
(sec)
0.4
1-7 x 106 Btu/hr
1.3 x 106 Btu/hr
0.85 x 10* Btu/hr
0.8
3.2
3.6
Figure 2-12. Effect of la stage residence time.
-------
additional benefit is gained past a residence time of about 1 second.
Similar curves were made for lower SR-, 's and the general trend is the
same as at this stochiometry. These curves were used to determine the
real effect of the la stage stoichiometry at a constant T, presented
earlier.
2.3.4 Residence Time, Ib Stage
The slow decay experienced in the la stage past a residence time
of 1 second appears to continue in the Ib stage as shown in Figure 2-13
at an SR^ of 0.80. This data is at a constant injection position so that
T, will be varying with SR-, and firing rate. However, at the third ter-
tiary air position, the residence time for most SR^'s and firing rates is
sufficiently long (>1.0) that the la stage residence time should not seri-
ously effect the results. It is thus interesting to note that the data at
a firing rate of 1.3 and 1.7 x 106 Btu/hr coincides at the same la stage
stoichiometric rato. However, at the lower firing rate of 0.85 x 106 Btu/
hr, the NO levels appear to be a bit lower. This could possibly be due
to the longer residence time in the la stage and/or due to a lower tempera-
ture environment. In the Ib stage at an SR,. of 0.95, the decay rate was
generally less than at an SR,. = 0.80 as illustrated in Figure 2-14.
In summary then, stack NO levels decayed both in the la stage and
the Ib stage. The decay appears to be fairly rapid in the initial 1 second
in the la stage, then drops off to a relatively slow decay in the Ib stage.
2.3.5 Firing Rate
In the previous sections on the effects of SR, , SR,b, T, and T,.,
the effect of firing rate has been discussed. Since firing rate effects
both local mixing, temperature and residence time between two given points,
it is often difficult to determine which of these effects is is predominant.
30
-------
480
440
400
360
320
280
CM
0 240
O
o
200
160
120
80
40
0
0.5
Western Kentucky coal
4 I FRF burners
B&W-type spreader, SW = 4
600°F preheat
15". excess air
Third tertiary air position,
SRlb = 0.8
O 0.85 x 106 Btu/hr
O 1 .3 x 106 Btu/hr
A 1-7 x 106 Btu/hr
I I I I I I I I I I I
0.7
0.9 1.1
1.3
1 .5
T-,, residence time (sec)
I h
SRla = 0.3
SR1a = 0.45
SRla =0.60
1.7
1.9
Figure 2-13. Effect of Ib stage residence time.
-------
CO
ro
600
500 -
400 -
300 -
200 -
100 -
Western Kentucky coal
4 IFRF burners
B&W spreader, SW = 4
_ 600°F preheat
15" excess air A
Third tertiary air position _
SRlb = 0.95
A_
u + 3
O
Fully shaded - SR = 0.3
Half shaded - SR - 0.45
Open - SR = 0.6
1 1 1 I 1 l
) 0.2 0.4 0.6 0.8 1.0 1.2
A
A'
O 0.85 x 106 Btu/hr
O 1 -3 x 106 Btu/hr
A 1.7 x 106 Btu/hr
l 1
1.2 1.4 l.f
Ib
Figure 2-14. Effect of Ib stage residence time.
-------
Thus, we see different effects depending on SR, and SR,. , or T, . However,
in summary, the only area where the attributes of firing rate are beneficial
is at the optimum SR, of 0.6 and a T, > 1.0 seconds with either an SR,K
i a i a i D
of 0.80 or 0.95. It is believed this is primarily due to a more rapid decay
rate in the la stage associated with higher temperature.
2.3.6 Effect of Temperature
Except for the resultant effect of temperature noted in the pre-
vious section due to firing rate, only a few tests were run with no pre-
heat to the secondary air. These tests run at a firing rate of 0.85 x
106 Btu/hr showed no significant effect. It is possible that the change
in secondary air temperature was not significant enough to change the com-
bustion temperatures.
2.3.7 Comparison with Conventional Staging
The question may be asked: Has anything really been gained by going
to this more complex staging arrangement? Table 2-5 will aid in answering
this question. Let's consider the optimum SR, = ~0.6 seconds in the la
stage with an SR^b of 0.95 at a residence time of 0.97 seconds. This yields
a stack NO level of about 400 ppm. The average stoichiometric ratio over
this time period is about 0.82. Now conventional staging at an SR = 0.95
yields an NO level of about 700 ppm. However, at an SR close to the average
for the total residence time of 1.57 seconds, 400 ppm is also achieved. Thus,
it appears that the distributed air concept has not really improved upon the
conventional staging result unless it is better to be at a very low SR for a
brief period followed by a higher SR for another time element as opposed to
being at the average stoichiometric ratio for the total time period. A simi-
lar conclusion is drawn at an SRla of 0.6 and SRlb of 0.80. In fact, it
33
-------
TABLE 2-5. DISTRIBUTED AIR VERSUS CONVENTIONAL STAGING
Arrangement Stage
Case 1
Distr. air la
Ib
Conventional staging
Conventional staging
Case 2
Distr. air la
Ib
Conventional staging
SR
0.6
0.95
avg. 0.82
0.95
0.85
0.6
0.8
avg. 0.73
0.75
T (sec)
0.6
0.95
Total 1.57
1.57
1.57
0.5
0.97
Total 1.57
1.57
NO (0% 02) ppm
400
700
400
300
250
34
-------
appears the conventional staging produces even lower NO levels than the
distributed arrangement for the same time period for this particular com-
bination.
2.4 CONCLUSIONS
In summary, a number of tests were conceived to explore a distributed
air concept to achieve low NOV emissions in a relatively short overall resi-
J\
dence time. This concept had proven successful in a premixed, small-scale
facility. Unfortunately, the results of the current study did not achieve
any improvement in the N0₯ time to staging results achieved with conventional
X
staging. This may be explained partly by the fact that a diffusion flame
was utilized in this experiment as opposed to a premixed flame in the smaller
scale experiment. Attempts were made to achieve as premix a situation as
possible by utilizing four burners and increasing the burner exit velocity
to effect a high mixing rate near the burner. However, the diameter of
the firebox (33 inches) results in a relatively low L/D for any given resi-
dence time, especially compared to the Pershing facility (Reference 2).
Table 2-6 compares these two facilities.
TABLE 2-6. FACILITY CHARACTERISTICS
Parameters
Diameters (in.)
Length (in.) for
~1 sec residence
time
L/D
EPA
Hor. Ext.
33
12
0.364
Pershing
6
54
9
35
-------
Thus, at the shorter bulk residence times in the EPA facility, it will be
difficult to achieve a real bulk residence time. That is, the real resi-
dence time, for residence times under 1 second, will be much less than the
bulk residence time due to a nonuniform velocity profile across the diam-
eter of the firebox.
It was found that with the distributed air approach, NO levels in-
crease with decreasing SR below an SR of about 0.6. In addition, NO always
decayed with increasing residence time in both the la and Ib stages.
In conclusion, no advantage was found for the distributed air con-
cept as applied to the diffusion burner arrangement in the EPA Multiburner
Facility.
36
-------
SECTION 3
COAL/OIL MIXTURE TESTS
As part of the Phase III alternate fuels testing, an emissions
evaluation test program was developed to look at coal/oil mixtures (COMs)
fired in a simulated package boiler configuration. Supplementing industrial
oil supplies with coal in the form of coal/oil mixtures has been investi-
gated for nearly 100 years (Reference 4). Over this period, the feasibility
of coal/oil technology has been demonstrated in both small-scale testing
and practical application.
With new interest in this technology, it is necessary to determine
if technology developed to minimize the environmental effects of coal com-
bustion is applicable to coal/oil systems or if further work is necessary
to ensure that pollution standards can be met.
OBJECTIVE
The purpose of this study was:
1. To obtain emission data for coal/oil combustion in an environment
closely simulating an industrial package boiler
2. To determine if emission levels were affected by the fuel com-
position
3. To determine if conventional control technology developed for
coal is effective in reducing emission levels produced by coal/
oil combustion
37
-------
4. To investigate the effect of burner modification on emission
levels produced by coal/oil combustion
Facility
This study was conducted in the EPA Multifuel Furnace Facility. The
experimental facility, as shown in Figure 3-1 and described in detail else-
where (Reference 1), was designed to simulate the aerodynamics of either a
front-wall fired or tangentially-fired boiler.
In order to simulate the heat release and temperature profiles con-
sistent with typical industrial package boilers, the additional modifica-
tions, as shown in Figure 3-2, were made. This configuration uses horizontal
extension sections that can be attached to the main firebox. These 33-inch
inside diameter by 6' long refractory-lined sections allowed the simulation
of a tunnel-fired unit and staging of combustion air at residence times
typical of what would be available in a package boiler. Additional hardware
included water tubes placed in the horizontal extension sections for addi-
tional heat absorption in the radiant section, and a heat exchanger placed
between the firebox and the horizontal extensions to achieve a gas tempera-
ture profile consistent with the radiant and convective sections in a typi-
cal industrial package boiler.
The burner used for the study was an IFRF 1.5 x 106 Btu/hr wall-
mounted unit. This burner is a versatile experimental swirl block burner
patterned after that developed by Beer (Reference 5). During baseline tests
on the parent coals, a Babcock and Wilcox-type coal spreader was used to
achieve a turbulent flame condition. Two commercial fuel oil atomization
nozzles were tested with the coal/oil mixture. These were the Delavan Cor-
poration swirl-air nozzle, shown in Figure 3-3, and the Sonic Development
Corporation Sonicore nozzle, shown in Figure 3-4. The burner was modified
38
-------
Combustion chamber (39" cube)
Ignition and flame safeguard
Observation ports
Ashpit
1.5 x 10* Btu/hr I FRF burner
C.E.-type corner fired burners
3200°F refractory
Heat exchange sections
Drawer assemblies
0. Staged Injection ports
Figure 3-1. Acurex/EPA multifuel furnace.
39
-------
O OJ
!'/
ft
'/
/ //I
/'' //i I/
/
i '/"/
{} ff
Existing Heat
Exchange Sections
R 3 ft ft
"' *f 'V /V
' 'f /I1 / <
/ll/<',J1/"
'"//I '//I1/'
'['/ '//'»' '
!«"<' '/'
A' " i/'
B tt 0 b
I FRF BURNER
V
Radiant Section
Convective
Section
V
Main Firebox
Figure 3-2. Facility modifications,
-------
MIXING CHAMBER
AIR INLET
FUEL INLET
AIR INLET TO
MIXING CHAMBER
EROSION
OCCURRED
IN THESE
AREAS
PINTLE PLATE
METERING NUT
Figure 3-3. Delavan swirl-air nozzle, 33373-1, 60 gph,
70° spray angle, mild steel construction.
41
-------
Fuel
IN5
Air
Resonator chamber
Erosion occurred
in these areas
Standing shock wave
Figure 3-4. Sonic Corporation Sonicore nozzle, 281T-B-11 with
stellite resonator chamber.
-------
for a low NOV configuration by placing an annul us of tertiary air around
A
the diffusor. This will be described later.
Emissions Monitoring Equipment
Continuous monitoring equipment was utilized to collect and record
data during this study. Table 3-1 lists the instrumentation used and the
principle of operation for each unit.
Fuel Preparation
The coal/oil mixtures examined in this study were prepared from par-
ent fuels which represent a broad range of classifications and fuel compo-
sitions. The compositional analyses of the parent fuel oils are listed in
Table 3-2 and that of the parent coals in Table 3-3. From these parent fuels,
four mixtures of 30 percent by weight coal to oil were prepared. Table 3-4
lists the mixtures and their compositional analyses.
The coals, pulverized to 70 percent through 200 mesh, were blended
with the fuel oils in a high turbulence batch mixer supplied by Littleford
Brothers. A suspension additive, supplied by Carbonoyl Co., was added to
ensure a homogeneous mixture. The additive was prepared as a 5-percent
aqueous solution and constituted a 3.75-percent by weight of the total mix-
ture. The mixture was prepared on a batch basis and 55-gallon drums were
used to store the fuel. The mixtures were stored at ambient temperatures
(50 to 60°F) for up to 21 days before they were fired.
Approximately 4 hours prior to use, each drum was wrapped with heating
blankets, and a mixer with a 6-inch propeller was immersed in the mixture.
The propeller was located approximately 6 inches from the drum bottom. The
mixture was heated and agitated utilizing a pump recirculation system until
the mixture temperature reached 150 to 170°F. The mixture was then pumped
43
-------
TABLE 3-1. EMISSION MONITORING EQUIPMENT
Pollutant
NO
so2
CO
coz
°2
Parti cul ate
Loading
Principal of
Operation
Chemiluminescence
Pulsed Fluorescent
Nondispersive
Infrared (NDIR)
Nondispersive
Infrared (NDIR)
Paramagnetic
Cyclone and
Filtration
Manufacturer
Ethyl Intertech
Thermoelectron
Ethyl Intertech
Ethyl Intertech
Ethyl Intertech
Acurex Corp
Models
Air Monitor-
ing
Teco
Model 40
Uras 2T
Uras 2T
Magnos 5A
HVSS
Instrument
Range
0-5 ppm
0-10
0-100
0-250
0-1000
0-5000
0-50 ppm
0-100
0-500
0-1000
0-5000
0-500 ppm
0-2000
«
0-5%
0-20*
0-5%
0-21%
0-3 pm
Minimum
-------
TABLE 3-2. FUEL OIL ANALYSES
Specifications^- "
^ ' Fuel Oil
API Gravity
Flashpoint, COC°F
Viscosity, SSU at 100°F
Heat of Combustion Btu/lb
Ultimate Analysis (% Wt)
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Amerada
Hess #6
15.3
204.0
2,500.0
19,867.0
84.71
10.75
0.36
1.93
2.22
0.03
Chevron
#6
12.3
182.0
4,900.0
18,292.0
85.57
10.52
0.81
2.08
0.93
0.09
-------
TABLE 3-3. COAL ANALYSES, AS-RECEIVED BASIS
Proximate (%Wtl_- "
^^_^ ^"""^ Coal
Moisture
Volatiles
Fixed Carbon
Ash
Rank
Ultimate (% Wt)
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Heat of Combustion, Btu/lb
Montana
21.23
35.16
34.27
9.34
Sub-bit. C.
53.26
3.35
0.87
11.16
0.78
9.34
8,972
Virginia
0.31
31.9
51.4
16.5
High-Vol. A
71.11
4.46
1.68
4.24
2.02
16.5
14,079
W. Kentucky
5.0
36.55
50.98
7.47
High-Vol. B
69.79
4.79
1.34
8.65
2.95
7.47
12,349
-------
TABLE 3-4. COAL/OIL MIXTURE ANALYSES, AS-RECEIVED BASIS, 30% COAL BY WEIGHT
^"~"-->^Jixture 30% (Wt;
^~~~"--^^^^ Coal
Analysis ^-^^^
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Moisture
Heat of Combustion,
Btu/lb
W. Kentucky/
Amerada
80.23
9.00
0.63
3.92
2.44
2.26
1.52
17,600
Montana/
Amerada
75.27
8.57
0.49
4.67
1.79
2.82
6.39
16,600
Montana/
Chevron
75.88
8.37
0.83
4.77
0.89
2.87
6.37
15,500
Virginia/
Chevron
81.23
8.70
1.07
2.73
1.26
4.95
0.09
17,000
-------
into the coal/oil mixture delivery system, shown schematically in Figure 3-5.
After each drum was emptied, the drums were inspected for settling of solids.
In all cases, little or no deposits were noted. The delivery system consists
of a 120-gallon capacity heated storage tank with an agitation system simi-
lar to the one described above. The fuel was kept well agitated and at 180
to 200°F. The mixture was delivered to the burner, through heat traced
lines and an electrical heater, via a Viking C-32 rotary pump with a vari-
speed control. A recirculation loop ensured that a homogeneous mixture was
maintained between periods of mixture firing.
A summary of the problem areas associated with this flow system
include the following:
a Coal settled out over a period of time in the bottom outlet of
the fuel holding tank causing complete plugging.
Plugging of lines in any low section of piping. Adequate
velocities must be maintained to keep the material entrained.
Shorting of electrical heat tape elements on piping and drums.
It is recommended that steam tracing be used if at all possible
in future tests.
Deterioration of pump performance due to wear at the seals and
increased clearances. Special seals for handling this very
abrasive mixture should be considered when pumping COM
An immersion heating element was used in the drums for initial
heating of the mixture before transfer to the holding tank. If
the mixture had settled these elements would not heat the mixture
uniformly and fires could easily develop. Also the tanks could
48
-------
HEAVY OIL
SOLENOID
VALVE
BURNER
NOZZLE
SAMPLE
LINE
VIKING
ROTARY
PUMP
With Varispeed
Control
ELECTRIC
HEATER
Fuel 3-5. Coal/oil delivery system.
-------
not be mixed successfully until they were thoroughly heated.
Thus it was found to be quite difficult to get these drums
reheated and well mixed after they had been setting for several
days.
The most difficult task was pumping the mixture from the drums
to the holding tank. The pump performance frequently deteriorated
to the point where it would not draw from the drum. This was
of course hampered by reheating the drum and getting the mate-
rial well entrained.
3.1 TEST PLAN
The tests were planned around a range of fuel types, three coals and
two oils, to determine if emission levels both under baseline and incor-
porating NO control technologies are dependent on fuel types. Table 3-5
A (
lists the baseline test matrix for the fuel combinations of interest.
Initially a 50 percent mixture of the various fuels was to be tested but
because of budget constraints and the mixing/handling problems encountered,
it was decided, with the project officer's concurrence, to limit the test-
ing to 30 percent concentration of coal. The baseline tests were run at 20,
30, and 40 percent excess air levels at a firing rate of 1.8 x 106 Btu/hr.
This firing rate gave a heat release per unit volume of about 50,000 Btu/hr-
ft3 which is typical of package boilers. The coal tests were run with a
B&W-type coal spreader with 15 percent primary air, and the coal and oil
tests were run with the Sonicore nozzle.
Table 3-6 shows the matrix for the effect of load and the effect of
residence time with staging as the NO control technique. All of these tests
A
were run at 20 percent excess air and 30 percent COM. Table 3-7 lists the
-------
TABLE 3-5. BASELINE MATRIX
fD
o
O
£
3
Q.
Coal /Oil
Mixture
r-
O
0)
i.
3
Q.
i
S
s
1.
5
1/1
CO
01
u
X
LLJ
O
^J-
s
o
o
CM
0
1-
O
CO
0
CM
Pure Fuels
Western
Kentucky Coal
22 5d
225c
225b
226g
226f
226e
6
§
>
01
£
O
217c
217a
217b
<5
Pennsylvania
222g
222f
222e
Coal /Oil Mixtures
Coals
Western
Kentucky
Montana
Virginia
Fuel Oils
Chev
PA
21 7d
217e
217f
Chev
2215
221 a
221 c
PA
221 r
221q
221n
Chev
218b
21 8a
21 8c
PA
51
-------
TABLE 3-6. EFFECT OF LOAD AND RESIDENCE TIME
O)
01
Residence Time - Short
0)
c
o
_1
1
QJ
*!
1
(U
U
c:
O)
-a
r
CO
OJ
OL
Tertiary
Distribution
O
CVJ
1
1
o:
GO
in
o>
o
1
oT~
oo
m
o>
o
1
oc.
oo
rO
oF~
CO
£
fO
E
£
Q_
in
00
0
in
r^«
o
in
UD
o
LT>
in
o
in
oo
o
in
i-^
o
in
10
o
in
in
o
in
00
o
m
r>-
o
in
ID
o
m
*d-
o
Coal /Oil Mixture 30%
Mont/Chev WKC/PA VA/Chev
Load x 106 Btu/hr
1.2
224£
224J
1.8
223f
223g
223h
223i
223o
223m
223k
223j
1.2
1.8
217k
1275,
217n
219o
219a
219b
21 9c
21 9d
1.2
1.8
220g
220h
220 i
220j
221 f
220j
220k
220£
52
-------
TABLE 3-7. DISTRIBUTED AIR BURNER TESTS
Ol
CTI
(O
-|J
l/l
Residence Time -Long
Residence Time -Short
oF~
CO
O)
O>
M
(/)
CO
r
in
o
in
CO
o
in
r-
o
if)
VD
O
in
-------
tests for the low NOV burner and combined low NO burner and staging con-
X X
figurations. The purpose of this matrix was to look at the effect of these
control technilogies on three fuel combinations. A range of the control
core stoichiometric ratios were tried with and without staging. Finally a
few tests were run by doping the fuels with pyridine and/or thiophene to
increase the nitrogen and sulfur levels appropriately. These tests are
shown in Table 3-8. A comparison of stack emissions with these dopants or
with fuels that naturally had these levels would then be possible. On all
of these matrices the test number has been given so that reference to the
emission levels may be determined from the listing in the appendix.
3.2 TEST DATA
The testing was divided into three phases. In Phase I the combustion
stability of each fuel was evaluated and delivery conditions were adjusted
for optimization of flame stability and combustion.
In Phase II of the study, some of the established combustion control
technologies for pulverized fuel were applied to the coal/oil mixtures. These
included staging of combustion air (Reference 6) and burner air distribution
(Reference 7). Flue gas recirculation, which has been found to be effective
in reducing thermal NO (Reference 6), was not applied due to equipment prob-
lems.
The last phase of the program was to evaluate fuel nitrogen conver-
sion by addition of dopant to the feed system. Results from each of these
phases will be discussed in the following sections. In addition, compari-
son will be made with data taken during a previous DOE-supported test in
this same facility. Details of the DOE tests may be found in the appendix.
54
-------
TABLE 3-8. EFFECT OF FUEL NITROGEN
u.
o
c
(U
O1
O
+J
s-s
\
CT» \
i \
X
\"
0 \
\°
o \
\ e=>
u\
o \
\5
0 \
% Sulfur (DMMF)*
0.9 ^^^
^^^ 1.2
Montana Coal
226a
Virginia/Chev
21 8c
Chevron Oil
217b
Mont/Chevron
221c
1.96
Mont/Amer + N
221o
Mont/Amer + N
221p
Mont/Amerada
221n
2.2 ^^^
^^^ 2.6
Virg. Coal
226e
Va/Chev + S
221k
Va/Chev + S
221m
W. Kent/
Amerada
217f
Amerada Oil
222e
3.4
W. Kent Coal
225b
DMMF: Dry, mineral matter free
55
-------
3.3 BASELINE TESTS
The baseline tests were designed to determine the optimum burner con-
ditions for each fuel and then the NO emissions as a function of excess air
for each parent fuel and fuel combinations. Initially the burner was
adjusted (swirl, axial fuel tube position and nozzle atomization rate) for
maximum flame stability for each fuel. In all cases, the nozzle position
was virtually the same relative to the IFRF burner, i.e., 2 inches forward
of the burner throat. However, atomizing air pressure had to be optimized
for each of the fuel mixtures. This was due to carbonaecous deposition or
"clinkering" on the water-cooled quarl in cases of poor atomization. In
general, the atomization pressure ranged from 12.0 to 22.0 psig depending on
the fuel. In most cases, this was 2.0 to 4.0 psig greater than the fuel
delivery pressure. Throughout the tests, secondary air was preheated to
300°F to enhance solids ignition in the relatively cold environment of the
radiant section. Burner swirl was optimized on the parent fuels and on
the coal/oil mixtures. A swirl of 5 on a scale of 8 was used for the base-
line coal tests, and a swirl of 0 (of 8) was used for both the oil and the
coal/oil mixtures throughout the testing. (A zero setting implies no swirl.)
Tests conducted earlier at the EPA/Acurex facility for design support of the
full scale DOE/Lorillard, Danville, Virginia, COM facility yielded a compari-
son of the two commercial nozzles described above. After approximately
3 hours of testing on 30-percent coal/oil mixture, utilizing the 440 hardened
stainless steel Delavan nozzle, significant erosion was observed. Areas of
erosion are shown in Figure 3-3. Equivalent testing with the stellite
Sonicore nozzle revealed similar erosion rates in the areas shown on Figure
3-4. However, the erosion had less impact on the atomization characteristics,
56
-------
flame stability and emission levels for the Sonicore nozzle. On this basis,
all subsequent coal/oil mixture tests were conducted with the Sonicore nozzle.
Figure 3-6 illustrates the results of the baseline emissions tests.
While CO, C02, S02, and NO data were taken during testing, NO data were
considered primarily and will be discussed here. Problems with the S02
analysis unit during testing rendered the data useful only on a relative
basis. Generally, CO and C02 levels which are quantitatively valid reflected
good combustion burnout in all cases except during staged combustion tests
at long first-stage residence times (1.5 to 2.0 seconds). During these
tests, CO levels rose to as high as 800 ppm (0-percent 02). Detailed emis-
sion levels for each test condition may be found in the appendix.
Figure 3-6(a) illustrates NO emission levels from the Chevron No. 6
oil base mixtures along with levels from the parent fuels. As is expected,
the NO levels for the mixtures fall in an intermediate range between the
parent fuels. The same data for the Amerada-based mixtures are illustrated
in Figure 3-6(b). In this case, though, emission levels for the mixtures
are very near those of the parent oil. Little contribution from the coal
is evidenced. Certainly, there are several possible mechanisms to which
this may be attributed. These mechanisms include effects due to sulfur in
the fuel, atomization characteristics of the oil, and the manner in which
the particular oil and coal volatilize. The volatilization rate of the par-
ticular oil surrounding the coal particles may also effect the fate of the
fuel nitrogen coming out of the coal. For example, Figure 3-7 shows the
boiling point curve for the Chevron and Amerada oils. This shows that the
Amerada has a much higher boiling point curve than the Chevron oil. The
rate at which the fuel N comes off the oil can also influence these results.
57
-------
14OO -,
1200 -
10OO -
i 800 -
600 -
400
1200 -
1000 -
(M
o
I
o"
800 ~
600 -
400 -
Load - 1.8 x 106 Btu/hr
Air preheat - 300° F
Q Chevron No. 6
Q Virginia/Chevron 30%
/\ Montana/Chevron 30%
Montana Coal
Virginia Coal
Chevron based mixtures
10 20 30
% EXCESS AIR
40
Load- 1.8x 106 Btu/hr
Air preheat » 300'F
Q W. Kentucky Coal
Q Montana Coal
/\ Montana/Amerada
O W. Kty/Amerada
Amerada No. 6
- Amerada based mixtures
200
I
10
I
20
% EXCESS AIR
I
30
- 40
Figure 3-6. Baseline emissions.
58
-------
90
OJ
o
s_
O)
Q.
OJ
>
'o
>
OJ
LO
l/l
80 -
70
60
50
40
30
20
10
2
3
4
5
6
7
8
9
10
East Coast
Middle East
Indonesian
Venezuelan Desulfurized
Gulf Coast
Venezuelan
Alaskan
Wilmington
Cali fornia
Amerada
Chevron
I
I
I
400
500
500
600|
600
700
8()0
700
Temperature
900
| 1000
800°K
,1100°F
Figure 3-7. Boiling point curves (Reference 8)
59
-------
Figure 3-8 shows the percent nitrogen evolved as a function of temperature.
These curves are shown compared to a variety of other oils (Reference 8).
As can be seen from the curve the Amerada oil is one of the more "refractory"
as far as N evolution is concerned. These mass and nitrogen evolution rates
could indeed play a role in performance of the various fuel combinations to
the NO control strategies.
At this point there is insufficient data to ascertain which of the
various mechanisms is causing this effect. But it is important to note that
all fuel combinations do not necessarily behave in the same manner. With
regard to the coal baseline tests, the differences between the Montana and
Western Kentucky coals are quite minimal and are probably within the error
band of the data. The Virginia coal, which consistently showed a higher NO
level than the Montana coal, is probably a real effect due to the higher
fuel nitrogen in the coal. The comparison of the Montana and Western Ken-
tucky coals differs from previous data (Reference 9). The previous data
showed the Montana coal to be consistently above the Western Kentucky coal
by about 5 to 10 percent. The main difference between the tests is that in
this test the temperatures are 300 to 400°F cooler. It is possible that
the volatilization rate of the Montana coal changes with the temperature,
and the Montana coal may have had a higher fuel nitrogen evolution rate in
the previous tests.
3.4 CONTROL TECHNOLOGY TESTS
The results of staged combustion of the coal-oil mixtures are illus-
trated in Figure 3-9. As is illustrated, the two Chevron no. 6 oil-based
mixtures responded almost identically to the removal of combustion air in
the first stage. However, the Amerada-based mixture showed little response
60
-------
30
o
A
O
O
D
O
0
2
3
4
5
6
7
8
9
10
O)
o
QJ
CL
TD
OJ
OJ
QJ
cn
O
20
10
East Coast
Middle East
Indonesian
Venezuelan Desulfurized
Gulf Coast
Venezuelan
Alaskan
WiImington
California
Amerada
Chevron
500
600 700 800 900
Temperature (°F)
1000
1100
Figure 3-8. Nitrogen evolution curves (Reference 8).
-------
Stoichiometry = 1.20
Load = 1.8 x 106 Btu/hr
1st Stage Res. time = 0.75 sec
800
30% Montana/Chev.
30% Virginia/Chev.
30% W. Kty/Amerada
600
rv>
CM
o
o.
o
400
200
I
0.65
0.75 0.85 0.95 1.05
First stage stoichiometric ratio
1.15 1.20
Figure 3-9. Staging emissions.
-------
(10 percent) in emissions reduction. It should be noted from Table 2, that
the Amerada no. 6 oil has a very high sulfur content relative to the Chevron
parent oil. This compositional difference could possibly have contributed
to the ineffectiveness of staging on the Amerada-based mixture NO emissions
(References 10 and 11) or it could again be due to the volatilization char-
acter of the fuels.
Following the staging tests, distribution of burner air was applied
to the mixtures. Figure 3-10 shows the distribution scheme schematically.
In these tests air which normally makes up a percentage of the secondary air
was injected through an annul us 7 inches radially from the burner throat cen-
terline. This was done in order to enrich the flame core where fuel-bound
nitrogen is evolved. The results of this testing are illustrated in Figure
3-11. Each mixture responded favorably to the air distribution, but it is
evident that the composition of each mixture leads to a unique emissions
curve under these conditions. In this case the Amerada oil-based mixture
performed about the same as conventional staging, that is showing little
effect to the control technique except when the core was made very fuel-rich.
A moderate effect is seen with the Virginia coal/Chevron oil and is again
very similar to the conventional staging result. However the Montana/Chevron
mixture shows a much more dramatic effect compared to conventional staging,
reducing the NO levels by a factor of three. The Montana coal also yielded
the lowest NO levels during staging tests in the main firebox during earlier
testing (Reference 1).
The curves in Figure 3-12 illustrate the results of applying burner
air distribution plus staging vs. straight burner air distribution. The
purpose of this comparison was to evaluate the mixture responses to further
63
-------
Stage
I air
if
Fuel tube
SR,
EA
~f / > i i 1 I T
\
Present IFRF burner with staging
Tertiary
cinnu 1 us te "
Fuel tube ^ SRla
///////f
-f
///////
SR,
Stage
r
EA
SR - Stoichiometric ratio
EA - Excess air
Modified IFRF burner for low NO
Figure 3-10. Burner air distribution.
64
-------
Stoichtometry = 1.20
Load = 1.8 x 106Btu/hr
800
600
M
o
I
I
o
z
400
cn
200
V 30% Montana/Chev
0 30% Virginia/Chev
Q 30% W. Kty/Aimrada No. 6
0.55
0.65
0.75
0.85
SR
1A
Figure 3-11. Burner air distribution.
-------
800 -
600 -
I -
o
2
200 -
Q SRj = 0.95
O SRi - 1.20
i i t \
0.55 0.65 0.75 0.85
FLAME CORE STOICHIOMETRIC RATIO
30% Virginia/Chev.
1st Stige Res. time » 0.75 tec.
15 x 106Btu/hr
SRi ' In Stage Stoichiometry
CN
O
Q
a
O
Z
800 -i
600
400 -
200 -
SR,
0.95
1.20
0.55 0.65 0.75 035
FLAME CORE STOICHIOMETRY
-------
enrich the flame zone. In Figure 12(b), for the Western Kentucky/Amerada
mixture, the further enriching of the flame zone resulted in higher NO
levels. This result, possibly, further illustrates the response of the
Amerada high-sulfur parent oil for fuel-rich conditions. The Montana/Chevron
mixture results are illustrated in Figure 12(c). This comparison validates
the interesting way in which each mixture reacted to the same combustion
conditions.
Again, we see little effect of the Amerada-based fuels, a moderate ef-
fect on the Virginia/Chevron mixture and a strong effect on the Montana/
Chevron mixture. It is interesting to note, however, that there was very
little difference between the air distribution or low NO burner tests and
X
the stated low NO burner tests on the Chevron-based mixtures. This may
indicate that with the short residence time staging there is considerable
backmixing of the stage air into the first stage. This backmixing would
result in a higher overall stoichiometric ratio in the first stage. How-
ever, the fact that the Amerada-based fuel reacted differently may indicate
there is more a dependence on the N and mass evolution rate and that the
timing of the distributed air may be unique for each fuel in order to achieve
an optimum low NO condition.
A
The effect of firing rate and residence time on NO emissions during
staged combustion is illustrated in Figure 3-13. Figure 3-13(a) represents
staging combustion air with a first-stage residence time of approximately
0.75 seconds. A first-stage residence time of approximately 1.50 seconds
is applicable to Figure 3-13(b). As expected in both cases, NO emission
levels decreased at the lower firing rate of 1.20 x 106 Btu/hr. NO levels
dropped to about one-half the baseline levels by increasing the residence
67
-------
800 -
600 -
CM
O
- 400 -
I
a
O
200 -
Q ljBx106Btufhr
1.2 x 106 Btu/hr
30% Monuni/Chev.
1ft Stage He*, time * 1.50 sec
1st Stage stoicriiorrwtry "1.20
0.65
0.75
I
0.85
SR,
B
0.95
1.05
I
1.20
30% Montana/Chevron
Stoichlometry 1.20
1st Stage Res. time 0.75 sec
is
O
O
00 -,
00 -J
400 -
too -
O
1* xlO6 Btu/hr
1.2x106 Btu/hr
0.86
0.75
I
0.85
SR,
1.05
1.20
Figure 3-13. Effect of firing rate and residence time.
68
-------
time. This result of residence time is coincident with earlier tests on
coal only that showed marked decreases in NO with residence time during
staged combustion. The levels at the longer residence time are similar to
the low NOX burner tests on the COM. However, CO levels during the long
residence time tests rose to 800 to 1000 ppm (0-percent 02) at times.
It should also be noted that the differences in NO at the two loads
decrease with decreasing first-stage stoichiometric ratio. This is again
consistent with previous coal data.
3.5 FUEL NITROGEN STUDIES
The last phase of the study was a limited evaluation of fuel nitro-
gen converion in COMs. This involved addition of a dopant to the feed
system to evaluate the conversion of fuel bound nitrogen. Pyridine, CgHgN,
was added to the Montana/Amerada mixture upstream of the fuel tube during
two tests, and the percent conversion of nitrogen to NO was calculated based
on the emissions levels obtained. These data, along with the emissions
levels of all the mixtures and parent fuels under baseline conditions, are
plotted against fuel nitrogen content in Figure 3-14. This plot exhibits
a definite trend. In both cases, the fuel nitrogen conversion to NO was
30 to 35 percent which is typical of bound-nitrogen conversion during coal
combustion. The pyridine dopant points fall slightly below the line, and
we have no definitive explanation for this. It is possible that the pyri-
dine volatilizes earlier than the oil or coal/oil mixtures, but this pre-
liminary screening test did not develop sufficient data to draw any defini-
tive conclusions. No correlation with fuel nitrogen was seen with the coal
data. This result is consistent with published data (Reference 8). It
appears that the COMs behave more like the oil with regard to fuel nitrogen
than to coal.
69
-------
1500
1300
1100
CL
CL
C\J
0 900
o
o
700
500
300
0.
O
1
1 .8 x 10 Btu/hr
20; EA
Sonicore nozzle
A
D
^ Montana/Amerada
^ Virginia/Chevron
0 Montana/Chevron
(2i W. Kentucky/Amerada
O W. Kentucky coal
A Virginia coal
O Montana coal
O Chevron oil
O Amerada oil
C^ Montana/Amerac- + dopant
O Montana/Amerada + dopant
0.50
0.70
0.90 1.10 1.30 1.50
Fuel nitrogen DMMF (^, wt)
1.70
1.90
2.10
Figure 3-14. Effect of fuel nitrogen.
-------
3.6 BURNER NOZZLE COMPARISON
NO emissions data were obtained with both commercial nozzles de-
scribed earlier on the parent oils. Figure 3-15 illustrates the results
of this comparison. Emissions data obtained earlier during the DOE tests
(see appendix for report on these tests) using the Delavan nozzle are
also included.
In all cases the Sonicore nozzle gave higher NO levels than the
Delavan nozzle. This result is attributable to the way the fuel is ato-
mized and mixed with the region. How a nozzle atomizes any given fuel can
also affect the NO levels. For example, the Chevron oil showed a more
marked difference between the two nozzles than did the Amerada oil. This
again shows the difficulty in trying to predict the NO levels for a given
oil and/or nozzle.
3.7 PREVIOUS TESTING DATA
During previous testing NO emission levels were obtained as a func-
tion of the percentage of coal in the mixture. These data are illustrated
in Figure 3-16. These data were obtained under baseline conditions with
the Delavan nozzle (hardened stainless steel) at a firing rate of 1.8 x 106
Btu/hr using the Virginia coal and Amerada parent oil. The results are not
too surprising for a given fuel and nozzle. They indicate that as the per-
centage of coal increases, the NO levels increase, although not quite
linearly.
3.8 SUMMARY AND CONCLUSIONS
The following conclusions can be drawn based on this study:
Baseline NO emissions from COM were, in general, linearly re-
lated to the fuel-N for a given burner and nozzle type
71
-------
800
600
LOAD: 1.8x 106 Btu/hr
SWIRL: Sonicore Oof 8
Delavan 3 of 8
X Amerada - Delavan DOE
V Amerada Sonicore
O Amerada - Delavan US
O Chevron Delavan
A Chevron Sonicore
ro
CM
o
g
o
400
200
20
30
40
% EXCESS AIR
Figure 3-15. Nozzle comparison.
-------
Q.
0.
CO
1400 r
1200
1000
800
600
400
200
Virginia/Amerada
Delavan Nozzle
1.8 x 106 Btu/hr
O 20% EXCESS AIR
A 30% EXCESS AIR
D 40% EXCESS AIR
40% EXCESS AIR
20% EXCESS AIR
10
OIL
ONLY
20 30 40 50 60
PERCENT COAL IN FUEL
70
80
90
100
PULVERIZED
COAL ONLY
Figure 3-16. Data from earlier work.
-------
The burner settings and fuel nozzle type have a strong effect
on NO emissions.
Conventional control technology utilized presently for pulverized
coal combustion is effective in reducing NO emissions but to
different degrees dependent on fuel composition.
- It was difficult to achieve low NO emissions with the Amerada
/\
based COMs.
- Moderate effects were obtained by NO control technologies
X
with the Virginia/Chevron mix.
Strong effects were obtained by NO control technologies with
J\
the Montana/Chevron. These were obtained with the low NO
/^
burner, low NO burner plus staging, and staging at long
A
residence times. However, CO levels are excessively high at
long residence times.
NO emissions increase in proportion to the amount of coal in the
coal/oil mixture but fall between the parent fuel oil and coal
baseline emissions.
Care must be exercised in designing the pumping systems for coal/
oil mixtures to avoid regions where coal may settle out and even-
tually plug the lines.
Flow control in small scale combustion tests of COM are quite
difficult due to the necessity of small orifices which can either
erode or become plugged.
Fuel nozzles which rely on impingement of solid surfaces by high
velocity fluid jets will be subject to high erosion rates. Judi-
cious selection of materials may help to overcome the problem.
74
-------
t Pumps must be selected which can handle this highly abrasive
mixture.
3.9 RECOMMENDATION
In order to analyze and understand the complex process of coal-oil
mixture combustion, we must obtain a better understanding of the combustion
processes of each parent fuel. Also, the combustion of coal/oil mixture
needs detailed examination to determine if it is merely a combination of
the two individual processes or if the interaction of these processes re-
sults in a completely different complexity of physical and chemical phenomena.
Future work should examine the role of fuel-bound nitrogen utilizing
flue gas recirculation and nitrogen evolution studies of the parent oils.
Also the effect of the physical presence that each fuel exerts on the other,
such as shielding or physical separation in the droplets, should be examined.
75
-------
SECTION 4
REFUSE-DERIVED FUEL TESTS
During the last decade, it has become apparent that the growing
demand for energy and the resulting scarcity of clean, readily available
sources to meet that demand have increased the need to use our energy re-
sources wisely and efficiently. The search for new fuels to supplement
present energy sources is taking on a new importance.
Commercial and municipal refuse has long been recognized as a vast,
untapped source of energy. However, the logistics of efficiently extract-
ing this energy has prevented its serious consideration as a viable source.
Dwindling supplies of clean fossil fuels have resulted in higher costs
along with restrictions in their use. While dirty fossil fuels are becoming
attractive alternatives, environmental considerations dictate that a balance
must exist in their use while new technology is developed to reduce their
harmful environmental effects. Alternative energy sources such as atomic
and solar are considered valuable but distant energy sources due to tech-
nological and environmental considerations.
In light of these considerations, fuel derived from refuse is a
practical energy source that is becoming increasingly attractive as tech-
nological advances overcome the problems inherent with its use. Fuel which
is derived from refuse by removal of noncombustible material has a nominal
76
-------
heating value of 4000 to 7000 Btu/lb. A significant number of investiga-
tions in which refuse-derived fuel (RDF) has been used to generate steam in
full-scale facilities have answered many questions regarding technological
problems associated with the use of this fuel. However, gaseous, trace
metal, and organic emissions data which can provide answers to environ-
mentally related questions are currently sparse.
It is necessary that investigations regarding the environmental
compatibility of refuse derived fuel be conducted before this vast, un-
tapped energy source can be considered a viable supplement to present en-
ergy resources. These investigations can be carried out most cost effec-
tively in a laboratory-scale facility.
4.1 OBJECTIVES
The objectives in this investigation were:
1. To design, fabricate and operate a system for combustion testing
of RDF in a laboratory-scale facility
2. To characterize RDF emissions of several types of material pres-
ently available for use as fuel
3. To evaluate the combustion efficiency of fuels consisting of
conventional clean and dirty fossil fuels (natural gas and coal)
mixed with varying percentages of RDF
4. To evaluate the effects of combustion parameters on emissions
from RDF/conventional fuel mixtures
5. To provide direction for future investigations on RDF to insure
solutions to problems associated with its use.
77
-------
4.2 RDF EXPERIMENTAL HARDWARE
In view of the fact that the majority of the U.S. steam-electric
capacity is produced by tangentially-fired boilers, all tests were conducted
in the tangential configuration. The EPA/Acurex multifuel facility, shown
in Figure 4-1, is capable of simulating several types of industrial and
utility boilers. For this investigation, the C.E.-type corner mounted burn-
ers, shown in Figure 4-2, were utilized. The aerodynamic pattern developed
in this configuration is shown schematically in Figure 4-3. The corner-type
axial diffusion burner shown in Figure 4-2 allows for gradual mixing of the
fuel and oxidant. In standard operation, fuel and primary transport air are
injected through the fuel tube of the center gun. A portion of the secondary
combustion air is injected through an annulus surrounding the fuel tube. The
balance of the secondary combustion air is injected through annular exits in
the upper and lower guns. Secondary fuel tubes, located in the upper and
lower guns, are used for system preheating with natural gas.
4.2.1 Burner Modifications
For this investigation, a fuel gun was designed to inject refuse
into the combustor and be aerodynamically consistent with the corner-type
burners. A modified gun, shown in Figure 4-4, was used for the injections
of the refuse material. The forward end of the nozzles includes an actively
cooled section to protect against preignition of the refuse. A thermocouple
was also installed on the nozzle outlet surface and connected to an over-
temperature alarm system.
The end of the nozzle was designed for a venturi effect to ensure
material transport velocities. An air injection port was installed at the
78
-------
1. Combustion chamber (39" cube)
2. Ignition and flame safeguard
3. Observation ports
4. Ashpit
5. 1.5 x 10s Btu/hr IFRF burner
6. C.E.-type corner fired burners
7. 3200°F refractory
3. Heat exchange sections
?. Drawer assembl ies
0. Staged injection ports
G
Figure 4-1. Furnace cross section.
79
-------
Yaw angle
Tangent
firing
circle
Heat exchange
section
RDF + air
Pri. fuel & air
Sec. air
Overhead view
Side view
Figure 4-2. Tangential configuration, aerodynamic patterns.
-------
Gas.
Secondary
air
Coal and
primary a1
Secondary
air
Secondary
ai
CO
Figure 4-3. Corner-fired burner.
-------
Intct
00
ro
\
\\ \\ v\ \\ \\
Water Cooling
With Active
Flow Path
Refuse
and
Primary
Air
-RJ *
Secondary Air
Figure 4-4. RDF nozzle.
-------
end to supply a portion of the combustion air. An access port was also in-
stalled to permit manual sweeping of material clogs with a rod if needed.
The modified gun assembly is shown in Figure 4-5. In the tangential
configuration, two modified burner assemblies were installed in opposite
corners, while two of the standard burner assemblies were installed in the
remaining opposite corners. This configuration is illustrated schemati-
cally in Figure 4-6.
4.2.2 RDF Feed System
A feed system specifically designed to eliminate problems associated
with the pneumatic transport of refuse material was fabricated for this
investigation. Handling problems, due to the nominal size of municipal
refuse (1/8" x 0 to 5.0" x 0), were significant at this scale of operation
(1.5 x 10 Btu/hr). A separate feed system was fabricated for each RDF
burner. Each unit was specifically designed to meter from 10 to 50 Ibs/hr
of RDF into the refuse nozzles described in the previous section. Signifi-
cant design effort was required to overcome many of the inherent problems
with feed refuse at these low flowrates. The system is shown schematically
in Figure 4-7. A review of the system follows.
The refuse feed system is made up of three main components which
were each designed to overcome specific handling problems involved with
refuse transport. These are:
Fluidized system
Belt transport system
Pneumatic transport system
Collectively, they result in a consistent feedrate of refuse into the com-
bustor.
83
-------
Refuse and
combustion air
Annular air
Coal & primary air
Gas
Secondary air
-
14 n LJ u u. a.
V \\ ", ^ v»
AiVWL
Cool ing
H20 lines
I t
Side view
Figure 4-5. Modified corner-burner assembly.
-------
Primary fuel
plus ai
00
on
RDF plus primary
fuel plus air
RDF plus primary
fuel plus air
Overhead view
Primary fuel
plus air
Figure 4-6. Fuel delivery schematic.
-------
Fluidization System. This system is made up of the refuse
supply drum, support frame, and a chain-driven motor system.
Refuse is placed into the chamber through a scalable door. An
electric motor is used to rotate the drum by way of a chain
drive/clutch system. The drum is rotated at a rate which ef-
fectively fills the drum with refuse in suspension, at steady
state conditions. This system results in an even distribution
of fluffy material which can easily be transported to the
burners.
Belt Transport System. An electric-motor-driven conveyor belt
system is located axially along the centerline of the drum. As
fluffed material falls along the moving belt, it is collected
by 1.0 inch metal stand-offs which are evenly distributed along
the belt surface area. The refuse layer thickness is controlled
by a pointed knife edge at the drum exit. The speed of the belt
is controlled at 0 to 1750 rpm by a varispeed motor.
Pneumatic Transport System. As the belt revolves around the
forward roller, the material is stripped off the belt using air
jets. The angle of these jets is such that they run tangent
to the roller, thereby effectively sweeping the belt clean of
material. This system also gives the material momentum into
the burner tube. The air used is part of the total combustion
air required for the refuse. The material is conveyed into the
furnace pneumatically through the RDF nozzle described previously.
A schematic of this system is shown in Figure 4-8.
86
-------
RDF feed system and firebox.
RDF feed system.
Figure 4-7. RDF feed system design.
87
-------
00
oo
Nitrogen
Supply
Compressed Air
Supply
Metal
Standoffs
Figure 4-8. Pneumatic transport system.
-------
Safety System
The flame safeguard system is designed to prevent flash-back or
propagation of refuse flames up the refuse injection tube and onto the belt
system. This possibility results from the positive pressure in the furnace
and the combustible mixture of air, refuse and dust in suspension in the
downcomer tubes. The system is shown schematically in Figure 4-9.
As shown, a Honeywell UV flame detector is positioned with a clear
line of sight into the burner tube. A flame signal from the detector is
sent to a control panel where three steps are taken automatically:
1. Refuse belt is shut off
2. Refuse air supply is shut off
3. Nitrogen purge is started
These steps ensure immediate loss of combustion essentials in the down-
comer and burner tubes. The activated nitrogen purge duration was set for
10 to 30 seconds.
4.2.3 Materials Acquisition and Handling
As stated in the Objectives section, one of the purposes of this
investigation was to better characterize combustion efficiency and emissions
as a function of refuse type. Therefore, material had to be transported
from several locations to the Acurex research facility. In order to assure
that the procedures used in obtaining this type of material complied with
state and federal regulations, coordination with local, state, and federal
authorities was required. In particular, coordination with the State of
California Department of Food and Agriculture was necessary because of
the potential entomological dangers of shipping RDF from various parts of
the country into the state. In order to protect against these dangers,
89
-------
Nitrogen
supply
Air
supply
NO
Mini peeper
UV scanner ^
Solenoid
valve
Plexiglass
tube
Burner tube
NO
Air
supply
Figure 4-9. Safety system.
-------
the State of California Department of Food and Agriculture, the Santa
Clara County Department of Agriculture, and Acurex agreed upon the follow-
ing packing, shipping and inspection procedures. The RDF material was:
Fumigated outside of the State of California using the procedure
recommended in the U.S. Department of Agriculture Plant Protec-
tion and Quarantine Manual, T403 (e)-(2), section 6, page 6.
This treatment is methyl bromide fumigation under a tarpaulin
using 10 pounds of methyl bromide per 1000 ft of RDF at atmos-
pheric pressure for 48 hours at 40°F or above.
Certified in writing by a state or federal agricultural agent
as to compliance with the fumigation procedure described above.
Sealed and shipped in a sturdy container with a rigid, insect-
proof frame and a leak-proof plastic liner. Wooden boxes, 44
inches x 48 inches x 100 inches, were constructed and fitted with
plastic liners for shipping the 2000-pound lots of RDF from each
each of the facilities.
Received and unopened until notification was given to the County
of Santa Clara Department of Agriculture so that a Santa Clara
County Agricultural agent could inspect the containers and the
written fumigation certification and agree to unloading.
After inspection and approval, the RDF was stored on the Acurex
premises in its shipping containers on an outdoor concrete pad and fully
covered by a tarpaulin. This protected the RDF from any degradation or
attrition from rain and wind during storage and testing.
Emphasis was placed upon safety during all RDF test operations. All
of the test personnel in contact with the RDF received tetanus, typhoid,
and diphtheria series of vaccinations, and personnel in direct contact with
91
-------
the RDF were required to wear respiratory filtermasks. In addition, daily
changes of clothing and footwear were required of test personnel in con-
tact with the RDF so that no contamination was carried around the work-
place or to the home. In addition, cleanliness of the test facility was
maintained. Floors and equipment were cleaned daily to insure a safe
and hygenic workplace.
The as-received RDF was not compatible with the test facility feed
system. Early calibration tests revealed that some of the large particles
in the RDF plugged the feeding mechanisms and interrupted testing. In
order to uniformly feed the RDF into the combustion test unit, the RDF had
to be reduced in size. Early testing with the test feeder system proved
that RDF particles of 1 inch or less were suitable for controlled combus-
tion testing.
Therefore, the RDF was passed through a commercial garbage composter
and reduced to 1 inch or less with no other alteration in the RDF composi-
tion. This size reduction process was conducted at the test site during
the combustion tests.
At the end of testing, any RDF which was unused was hauled to a
landfill area for disposal. The major consideration was to dispose of
the unused RDF as soon as possible to prevent insect infestation or putre-
faction of the RDF.
4.2.4 Sampling Equipment and Procedures
The sampling required for this project included collection of gaseous
emissions by continuous monitoring equipment, collection of flue particu-
late and gases for trace metal, organic and anion analysis, collection of
residual ash for detailed analysis, and sampling of the input fuel. The
92
-------
methods and equipment for each of these sampling tasks are discussed in the
following sections.
4.2.4.1 Gaseous Emissions Measurement
Table 4-1 lists the continuous monitoring equipments utilized at the
Acurex Energy Laboratory. Figure 4-10 shows a schematic of the gaseous
sampling and analysis system. The system is designed for accurate analysis
of NO, CO, 02, C02, S02, and unburned hydrocarbons.
4.2.4.2 Stack Sampling Equipment
Two stack sampling systems were used during the course of testing.
The high volume stack sampling system as shown in Figure 4-11 was used to
determine the particulate grain loadings and size distribution. This sys-
tem meets or exceeds the EPA Method 5 requirements. The second system was
the Source Assessment Sampling System (SASS) shown schematically in Figure
4-12. The SASS is used for both sampling of particulates and organics. The
sample is drawn through a glass-lined sampling probe and routed through a
series of three cyclones and a filter which separates the particulate into
four size fractions. Both the probe and particulate removal system are in
a 400°F oven to prevent condensation. The gaseous sample then passes
through an organic module where it is cooled and the organics are trapped
on a polymer adsorbent. Condensate from the module is also collected for
analysis. Finally, the sample is routed through an impinger train where
oxidizing solutions retain any remaining sample. The sample is then drawn
through the control unit where pressures, temperatures, and gas volume are
monitored and controlled. An S-type pitot is used to measure gas velocity
for the purpose of determining isokinetic sampling rates.
93
-------
TABLE 4-1. EMISSION MONITORING EQUIPMENT
Pollutant
NO
so2
CO
co2
°2
UHC
Participate
Loading
Principal of
Operation
Chemi luminescence
Pulsed Fluorescent
Nondispersive
Infrared (NDIR)
Nondispersive
Infrared (NDIR)
Paramagnetic
Flame lonization
Cyclone and
Filtration
Manufacturer
Ethyl Intertech
Thermoelectron
Ethyl Intertech
Ethyl Intertech
Ethyl Intertech
Ethyl Intertech
Acurex Corp
Models
Air Monitor-
ing
Teco
Model 40
Uras 2T
Uras 21
Magnos 5A
FID
HVSS
Instrument
Range
0-5 ppm
0-10
0-100
0-250
0-1000
0-5000
0-50 ppm
0-100
0-500
0-1000
0-5000
0-500 ppm
0-2000
0-5%
0-20%
0-5*
0-21%
0-100 ppm
0-300 ppm
0-1000 ppm
0-3 vim
Minimum
94
-------
Sampling probe heated filter
vo
en
filters Pulsed fluorescent
filter
Figure 4-10. Sampling system online at experimental multiburner furnace.
-------
unit
25-Foot umbilical line
Oven with
cyclone
and filter
Impinger train
and ice bath
25-Foot sample hose
0 cfm vacuum pump
Pump-control unit hose
Figure 4-11. Aerotherm high volume stack sampler.
-------
Fan
Oven
P1H«r
Sttel T.C.
^ Oven Fine
Sensor
umbilical
AP. T
Gas conditioner I
moisture col lector
Porous polymer
adsorber
IMtot AP
Gage
Parallel
Imp/cooler
trace element
col lectors
Dry Gas Meter
Orifice Meter
Centralized Temperature
and Pressure Rea.1 jt
Semiautomatic
control
module
Control Module
10 cfm vacuum pump
Figure 4-12. Source assessment sampling system (SASS).
-------
4.2.5 Problem Area Summary
As was discussed earlier, one of the objectives of this investiga-
tion was to design, fabricate, operate, and evaluate a laboratory-scale
system for combustion testing of RDF. Since, to our knowledge, no other
refuse investigation had been performed on this scale, it is our intention
to fully document the areas where problems occurred during this investiga-
tion.
Since the investigation was conducted in the suspension fired, tan-
gential configuration, as descrbed in Section 4.2.2, two complete refuse
systems were required for opposed refuse input. The complexity of each
system alone required careful monitoring along with the primary fuel systems.
The RDF feed systems were evaluated as follows.
Drum System
The mechanical fluidization system performed very well throughout
testing, although there were several times the system went down on one side
or the other. Problems resulting in downtime were:
1. The weight of the drum system caused compression of the forward
cam system and resulted in binding at the teflon bearing-space
interface. This problem occurred six to eight times during the
3-week test period and was quickly corrected each time by adjust-
ing the cam vertical position. This problem may be averted al-
together by utilizing a noncompressing material for the cam.
2. The nature of the refuse was such that small pieces of paper,
glass, etc., being fluidized in the rotating drum found their
way into the space between the nylon bearing seals in the for-
ward drum. This resulted in binding at least three times during
98
-------
testing. The problem was resolved by removing the seal cover
and cleaning the entire surface. This required 1 to 1-1/2 hours
to complete.
Fluidization of the material was excellent throughout testing as
long as the material was in a fluffy, dry condition.
Belt System
As was mentioned earlier, fluidization of the material was poor
when it was moist or packed prior to loading in the drums. This resulted
in clumping on the belt system which caused inconsistent input to the
burners. The drum exit was such that material agglomeration resulted in
build up at the exit. This periodically stopped the belt or caused the
belt to be stripped of material at that point.
A more efficient deflector design could solve this problem, but
more important is providing a properly conditioned feed consistently.
The feed was determined by calibrating the belt rpm against mass
delivered. While this system was somewhat accurate initially, as testing
on each fuel continued, it became apparent that the calibrations weren't
holding. Therefore, the RDF input was determined by back calculating from
the flue gas analysis. The errors resulted from the differences in refuse
density from layer to layer in the storage bins. Visible differences of
the refuse characteristics were evident from day to day on the same fuel
types.
Belt to Burner Transition
Most of the plugging problems incurred in this area were due to in-
consistency in the refuse sizing. When plugs occurred, sweeping the tube
clean was easy and quick, if approached properly. After initial trial and
99
-------
error, elimination of refuse plugs was a secondary problem. However, if
material flow sensors allowed buildup on top of the plug to occur, the
tube had to be removed in order to sweep the plug.
Generally, a coated or more scratch resistent tube is the only im-
provement that could have been made.
Plugging in the gun occurred when compacted material from the down-
comer was forced into the gun. Other plugs resulted from foreign objects
such as wire becoming lodged within the system.
Material Preparation
Generally, a great deal of the handling problems would be eliminated
if a material preparation system yielded the same type product each batch.
The problems resulting from this are probably nonexistent on a large scale,
but become relevant on this small scale. The material is very absorbent
in this state and should be guarded from heat and humidity.
Stack Sampling
The physical nature of the particulate product of RDF/gas cofiring
resulted in extended sampling periods to collect the required volume of flue
gas sample. As shown in Figure 4-12, all particles less than 1.0 micron
sizing were collected in a fiberglass filter upstream of the gas conditioner.
In all cases with RDF/gas cofiring, all solids collected were smaller than
1.0 micron. Therefore, frequent filter changes were required to complete
sampling.
During RDF/coal cofiring tests, the small particulate apparently
adhered to the larger coal ash particles and were captured in the cyclones.
This eliminated 80 percent of the sample and were captured in the cofired
tests.
100
-------
Ash Deposition
During RDF/gas cofiring, bottom ash deposition was minimal, although
after approximately 20 hours of testing at concentrations of 30 percent
refuse, on a heat input basis, some ash deposits were removed from the
ashpit.
The ash deposition during RDF/coal cofiring displayed characteris-
tics unique to that mixture. Daily ash collection was necessary for all
ratios of coal/RDF. However, during approximately 25 hours of testing at
a refuse concentration of 30 percent, on a heat input basis, a bridging
of ash occurred across the ashpit entrance. This occurrence is illustrated
schematically in Figure 4-13. The fused ash material recovered weighed ap-
proximately 100 pounds.
This problem has never occurred at this facility before during over
3 years of coal testing at somewhat higher ash input rates.
4.3 TEST THEORY AND PLAN
In order to achieve the objectives noted in Section 4.1, the experi-
mental program consisted of two basic elements:
Baseline tests -- Evaluation of feed systems and characteriza-
tion of emissions during gas/refuse cofiring.
Detailed tests -- Evaluation of combustion efficiency and conven-
tional emissions controls during coal/refuse cofiring.
4.3.1 Baseline Tests
As stated above, the purpose of the baseline tests was to evaluate
the refuse feed systems, using all four types of refuse, and characterize
the combustion performance and emission levels from each of the materials
cofired with gas. Based on the results of these tests, one of the refuse
materials was selected for use as the detail test fuel. Also, modifications
101
-------
Firebox exit
Ash bridge
Ashpit
entrance
Internal firebox, side view
Figure 4-13. Ash deposition.
102
-------
and adjustments of the refuse feed system could be made in order to assure
consistent fuel input during the lengthy detailed test points. The emissions
produced by each of the refuse types were used to illustrate the uniqueness
of each refuse and to obtain background values for the chosen detail test
fuel.
The test matrix developed for the baseline testing is shown in Fig-
ure 4-14. As noted on all the matrices, sampling is divided in three
levels of detail. Level 1 consisted of gaseous emissions sampling only.
Level 2 sampling included gaseous emissions and stack particulate loading
tests. The Level 3 sampling included tests under Levels 1 and 2 plus de-
tailed stack sampling for trace metals and organic compounds in the stack
flue gas.
As indicated in Figure 4-14, the bulk of the testing was completed
at 20 percent excess air conditions at a heat input rate of 1.5 x 10 Btu/hr.
Tests at other conditions were necessary for background levels to be used
with the detailed test results.
4.3.2 Detailed Tests
After completion of the baseline tests and selection of the detail
test refuse, the detailed test matrix, shown in Figure 4-15, was addressed.
As noted, the purpose of these tests was to evaluate the combustion effi-
ciency of a refuse/coal fuel mixture and to evaluate conventional emissions
control, i.e., theoretical air, on the resulting emissions. Sampling test
nomenclature was consistent with that used during the baseline tests.
As noted in Section 4.2.5, the particle size of the stack particulate
produced during the gas cofired points was such that sampling time required
for both Level 2 and 3 tests was increased by a factor of 2. This resulted
103
-------
4 RDF Types
1) Gas coflre
2) Theoretical air
3) S RDF
4) Residence time to convectlve section
5) Firing rate
TA - theoretical air; LRT - long residence time; SRT - short residence time;
105X. 110X. IZOt. 130t
St. lot, 201
(short, long)
1.0 x 10* and 1.5 x 10* Btu/hr
.1. Gaseous emissions sampling only
2. Gaseous emissions plus flue gas
participate loading and size
distribution
3. Detailed emissions sampling
u.
g
X
in
u.
o
Of
t*
o
u.
o
Of.
»
o
VI
fc
i/>
£
&
v\
&
_l
&
VI
t
_l
105S TA
1.0 x 10* Btu/hr
1.5 x 10§ Btu/hr
1
1
1
not TA
1.0 x 10* Btu/hr
1.5 x 10« Btu/hr
1
1
1
1201 TA
1.0 x 10* Btu/hr
1
1.5 x 10* Btu/hr
2
3
2
130* TA
1.0 x 10* Btu/hr
1.5 x 10* Btu/hr
1
1
1
Figure 4-14. Test matrix for baseline emissions characterization.
-------
1) Cofire RDF *1th coal
2) Firing r«te 1.5 x 10' Btu/hr and 1.0 x 10' Btu/hr
3) Vary residence tine to corrective section
4) U.C. RDF, variable X
5) Vary theoretical air
LL.
O
at
»«
ttt
u.
a
tr
S
Li.
O
a
»*
o
2*
§S5
_l CC »-
4J -D
l«l
VI -
0,2*
c M e
O «J ^
-ICC H-
*> "d
U f- «j
0 m E
^ OJ --
irt a t-
T>
cn-~- u
§vi E
D
ICK t-
t2fc
isl
Irt K t-
0,2 .
gjcl
-IQCt-
***
J88J1
lAOCI-
1051
1 n MBtu
'° hr.
1
7A
, c HBtu
KS hr
1
1
3
2
110S
, .. MBtu
1'° hr
1
1
1
, c MBtu
'5 hr
1
3
2
1
120J
1 0 MEtu
'° hr
1
2
2
1
TA
, c KBtu
K5 hr
2
3
3
2
130%
, n KEtu
'° hr
1
3
TA
1 5 M"J
'5 hr
1
3
1
2
1. Gaseous emissions sampling only
2. Gaseous emissions sampling plus flue gas
partlculate loading and size distribution
3. Detailed emissions sampling
Figure 4-15.
Test matrix for emissions control through
theoretical air variation.
105
-------
in a loss of sampling points during the detail testing. Figures 4-16 and
4-17 illustrate the completed test matrices. The detailed testing was
focused on Level 2 and 3 points as indicated.
4.4 ANALYTICAL PROCEDURES
This investigation required detailed chemical analyses of fuel
samples, and gaseous and solid stack product samples. All of these analy-
ses were completed at the Acurex Analytical Laboratory with the exception
of analysis of fuel samples.
The methodology of these analyses is outlined below.
4.4.1 Fuel Sample Analysis
Representative samples of all fuels tested during this investigation
were submitted to a certified commercial laboratory for ASTM standard
analyses listed in Table 4-2.
TABLE 4-2. FUEL ANALYSES
Proximate Analysis
Moisture
Ash
Fixed Carbon
Volatile Matter
Ultimate Analysis
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Chlorine
Heating Value
4.4.2 Trace Metal Analysis
Trace analyses of metals were conducted using atomic absorption
spectroscopy by standard EPA and ASTM methods. The metals which were ana-
lyzed are listed in Table 4-3. Particulate fractions from the sampling train
106
-------
4 RDF Types
1) Gas coflre
2) Theoretical air
3) 1 RDF
4) Residence time to convectlve section
5) Firing rate
105X, 110*, 120S, 1301
5t, 101, 201
(short, long)
1.0 x 10' and 1.5 x 10* Btu/hr
TA theoretical air; LRT - long residence time; SRT - short residence time;
.1. Gaseous emissions sampling only
2. Gaseous emissions plus flue gas
partlculate loading and size
distribution
3. Detailed emissions sampling
u.
o
Cf
M
in
i*.
a
Of
M
o
u.
O
ac
*
o
CM
&
vt
s
_)
f-
ae
i/i
cc
_j
S
i/i
fc
1051 TA
7.0 x 10* Btu/hr
1.5 x 10* Btu/hr
©
©
©
1101 TA
1.0 x 10* Btu/hr
1.5 x 10: Btu/hr
©
©
©
1201 TA
1.0 x 10* Btu/hr
©
1.5 x 70* Btu/hr
©
©
©
130t TA
1.0 x 10* Btu/hr
1.5 x 10* Btu/hr
©
©
©
O Completed tests
Figure 4-16. Test matrix for baseline emissions characterization.
-------
1) Coflre RDF *1th coal
2) Firing rate 1.5 x 10' Btu/hr and 1.0 x 10( Btu/hr
3) Vary residence time to convectlve section
4} U.C. RDF, variable X
5) Vary theoretical air
Li.
O
oc
X
U_
0
tt
o
0
o
CM
u.
0
ct
V*
o
c
(7>^- 0>
C wl £
O OJ
_J CC t
1- Ci
o >^ e
^: c.
uo ce »-
o
cm- o>
c m E
.
L_ ^ CJ
O on E
XT Qj ^~
T3
C l/l E
00
_4 CC t
«- XI
t-- y
O f» E
to tt; h-
o> w
C *^l t-
_i ce i
.
fe ₯
O kn 6
£ «> »
to et t
1051 TA
1 0 HBtu
'° hr.
1
, MBtu
1-5 hr
©
©
3
2
1101
l n HBtu
'° hr
1
1
1
, r MBtU
hr
©
©
©
©
120% TA
l n MEtu
'° hr
1
2
2
1
, ^ H3tu
" hr
2
-
©
©
©
120X TA
K5.tu
hr
1
3
i
, r M3tJ
''* hr
1
3
©
©
1. Gaseous emissions sampling only
2. Gaseous emissions sampling plus flue gas
partlculate loading and size distribution
3. Detailed emissions sampling
O Completed tests
Figure 4-17.
Test matrix for emissions control through
theoretical air variation.
108
-------
were analyzed after acid or Parr digestion. For each SASS train, at least
three samples were analyzed a proportionally combined representative
participate sample, a sample of the XAD-2 resin, and combined aqueous
condensate and first impinger solutions after extraction, and the combined
second and third impinger solutions. However, only antimony, mercury, and
arsenic were analyzed in the second and third impinger samples.
TABLE 4-3. METALS WHICH WERE ANALYZED
Trace Metals
As
Be
Cd
Hg
Ti
Sb
Sn
Pb
Cu
Mn
4.4.3 Organic Analysis
Organic species were analyzed by a modified Level 1 analysis scheme
(Level 1 Environmental Assessment, IERL-RTP Procedures Manual, June 1978).
Basically, this scheme involves the separation of a sample extract into
broad classes based on liquid chromatography fractionation and gravimetric
analysis. An organic extract is placed on a column of silica gel and frac-
tionated by elution with increasingly polar solvents. Table 4-4 lists the
solvents which are used in the Level 1 scheme. Each fraction after solvent
removal is weighed to yield a rough estimate of material present. This
separation scheme yields seven fractions which will contain the compound
classes outlined in Table 4-5.
Selected fractions from the liquid chromatography separation were
then scrutinized for specific chemical species. For this investigation,
109
-------
TABLE 4-4. LIQUID CHROMATOGRAPHY ELUTION SEQUENCE
Fraction
1
2
3
4
5
6
7
Solvent Composition
Pentane
20 percent methylene chloride in pentane
50 percent methylene chloride in pentane
Methylene chloride
5 percent methanol in methylene chloride
20 percent methanol in methylene chloride
50 percent methanol in methylene chloride
Volume
25 ml
10 ml
10 ml
10 ml
10 ml
10 ml
10 ml
110
-------
TABLE 4-5. DISTRIBUTION OF COMPOUND CLASSES IN LIQUID
CHROMATOGRAPHIC FRACTIONS OF ORGANIC EXTRACTS
Fraction Compound Class
1 Aliphatic hydrocarbons
Halogenated aliphatics
2,3 Aromatic hydrocarbons
Halogenated aromatics (PCB's)
4,5 Nonpolar oxygen or nitrogen
containing species
6,7 Polar compounds - phenols,
alcohols, amines, etc.
Ill
-------
the organic compounds of interest are prevalent only in the LC fractions
2 and 3. Therefore, samples were collected only from these fractions.
The sample was then analyzed by gas chromatographic/mass spectrometry
methods. During this analysis the level of investigation was determined
quantitatively utilizing the threshold level for nearly all the most toxic
species as defined by OSHA, that level being 0.10 mg/m of sample gas.
All peaks above this level were analyzed for the following groups or
species:
1. ROM's (polycyclic organic materials)
2. PCB's (polychlorinated biphenals)
3. Four other groups or species
The other groups or species were selected based on the largest
quantities of materials which did not fall into the two groups specifically
selected above.
4.4.4 Quality Assurance and Control
To assure the quality of the analytical data, a program used to
control contamination, calibrate instrument response, and verify qualita-
tive and quantitative data is presented below.
Glassware
All glassware used in the extraction and analysis of the samples
was cleaned by one of two methods. Separatory funnels and volumetric
glassware were cleaned in a dichromate acid bath, rinsed with deipnized
water, rinsed with acetone, hexane and methylene chloride, and sealed with
muffled aluminum foil. All other glassware was washed with soap and water,
rinsed with deionized water, rinsed with acetone and muffled at 450°C to
500°C for approximately 6 hours. Although not adopted as a standard
112
-------
procedure, this procedure has been used by Acurex and EPA labs to produce
glassware totally free of detectable organic contaminants for several years.
Solvents and Standards
Only Burdick and Jackson "Distilled in Glass" solvents were used
in this program. Acurex purchsed all solvents in lot quantities to assure
uniform quality throughout the entire study. A quality check was performed
on each solvent to insure the absence of any interfering substances prior
to the start of the program.
All standards were purchased from commercial supply houses or from
EPA. Each standard was verified by GC/MS prior to its use.
Blanks and Spikes
Two types of blanks were taken: (1) a sampling train blank for each
test and (2) method blanks for the solvent extractions. For each series
of test runs, a blank train was set up in the same manner as the actual
operating train. The blank train was capped off at the nozzle and impinger
exit with aluminum foil. The train remained assembled at the test loca-
tion for the duration of the test period. Sample recovery and analysis
proceeded as described for the sampling train. Method blanks using the
same glassware and solvents as for the actual samples were taken every 10
samples and analyzed as described earlier.
Metals Analysis
Trace metal analysis requires a careful adherence to good analytical
techniques and the measurement of spiked samples. To this end, each sample
was spiked to give an increase in the initial concentration greater than
10 percent but less than 100 percent. The recovery was calculated from
these data and applied to the values found.
113
-------
Standards were diluted from stock each day and a standard curve
plotted at the beginning and end of each analysis for that element. The
standard curve was selected in such a way as to bracket all of the sample
concentrations for the run. After each 10 samples, at least one standard
was rerun at the level that approximated most of the sample concentrations.
Replicates were run at regular intervals to establish precision of the method
and spike, and recovery for the accuracy data.
4.5 EXPERIMENTAL DATA
In this section, the experimental results for completed tests will
be presented. This will include data on the fuel samples, gaseous emissions,
particulate emissions, trace metals, and the organic emissions. Table 4-6
lists the test point designations and their corresponding test conditions
for referral from the test data.
4.5.1 Fuel Samples
During the testing phase of this investigation, the fuels were being
continually sampled to better characterize the inputs. At the completion
of the testing, these gross samples were combined and sent to a commercial
laboratory. Representative samples were drawn and analyzed as discussed
in the previous section. The results of those analyses are listed in Table
4-7. Photos of the fuel samples are shown in Figure 4-18.
4.5.2 Gaseous Emissions
While the objective of this investigation was primarily to charac-
terize organic and trace metal emissions from conventional fuel/refuse
fuel mixtures, gaseous emissions were also fully documented. Discussion
of gaseous emissions will be limited to oxides of nitrogen and sulfur
dioxide primarily, due to their importance in environmental considerations.
Full gaseous data are documented in the appendix.
114
-------
TABLE 4-6. TEST MATRIX
Test
Point
11A
B
C
D
13A
B
C
D
40
15
37
38
32
31
19
35
34
Fuel
Gas/Ames
Gas/Richmond
Gas/Americology
Gas/San Diego
Gas/Ames
Gas/Richmond
Gas/Americology
Gas/San Diego
Pitts Coal
Coal/Richmond
Coal/Richmond
RDF
Cone*
10%
10%
20%
20%
5%
10%
10%
20%
20%
30%
30%
Combustion Conditions
1.5xl06Btu/hr 20% EA
20% EA
10% EA
10% EA
20% EA
10% EA
20% EA
10% EA
20% EA
1.5xl06Btu/hr 30% EA
Heat input basis
115
-------
TABLE 4-7. FUEL ANALYSES
cr>
Ultimate Analysis*
Carbon %
Hydrogen %
Oxyen %
Nitrogen %
Sulfur %
Ash %
Moisture %
(as received)
Chlori ne %
Heating Value
Btu/lb
Fuel Type
Pittsburg
No. 8 coal
75.23
5.15
8.12
1.49
2.51
7.50
0.93
0.14
13,545
Richmond
refuse
42.60
6.26
37.90
0.83
0.16
12.25
23.8
.46
7696
Ames
refuse
40.49
6.01
30.04
0.73
0.35
22.38
15.2
.43
7R31
Ampri cology
refuse
40.29
5.88
25.20
0.91
0.17
27.55
24.4
.72
7164
San Dieqo
refuse
38.01
5.64
17.40
0.69
0.21
38.05
26.3
.79
7146
Dry basis
-------
Ames
Americology
Figure 4-18. Photographs of fuel samples
117
-------
/
Richmond
,*
'
San Diego
Figure 4-18. Concluded,
118
-------
It should be noted that during this and a previous investigation,
sulfur dioxide emissions data were inconsistent. Following this investi-
gation, the Pulsed Florescent S02 Analyzer was returned to the manufacturer
for evaluation. The source of the inconsistent data was determined to be a
photomultiplier tube which rendered the S02 data during this investigation
invalid on a quantitative basis. However, the data is valid on a relative
basis and should be regarded as such.
As discussed in the test plan, baseline testing to characterize the
combustion of refuse was conducted first. This was accomplished by co-
firing each of the four refuse types with natural gas and by examining
several variables. These variables included excess air and concentration
of refuse on a heat input basis. All other combustion parameters were
held constant.
The results of this baseline testing are illustrated in Figures 4-19
through 4-22 where NO is plotted as a function of excess air percentage for
each of the four refuse types. The refuse concentration effects are also
illustrated. In each figure, a baseline point is plotted. This point,
taken with natural gas as the fuel, represents NO formed through thermal
fixation of the atmospheric bound nitrogen. Figure 4-23 represents data
taken during previous work on natural gas. The baseline point taken during
this investigation is plotted to demonstrate the validity of the NO level.
Using this as a baseline illustrates qualitatively the contribution of fuel
bound nitrogen to the total NO emission.
It should be noted in Figure 4-20 that the 20 percent Richmond curve
falls below the curve representing 10 percent Richmond fuel. This is
119
-------
200
CM
O
^ 100
no
o
10 20
Excess air (percent)
30
O Gas Only
0 Ames 5%
D Ames 10%
O Ames 20%
(Heat Input basis)
Tangential mode
1.50 x 106 BTu/hr
300°F Secondary A1r
Figure 4-19. NO emissions during baseline testing (Ames)
-------
r>o
200
CM
O
V!
O
S- 100
10
O Gas Only
0 Richmond 5%
Q Richmond 10%
Richmond 20%
Tangential Mode
1.50 x 106 BTu/hr
300°F Secondary A1r
Excess air (percent)
Figure 4-20. NO emissions during baseline testing (Richmond)
-------
200
CVI
o
O.
Q.
100
IXi
I
I
10 20 30
Excess air (percent)
Gas Only
Q Americology
O Americology 20%
(Heat Input basis)
Tangential Mode
1.50 x 106 BTu/hr
300°F Secondary A1r
Figure 4-21. NO emissions during baseline testing (Americology)
-------
200
CM
O
Q.
Q.
100
ro
oo
10 20
Excess air (percent)
30
Gas Only
D San Diego 10%
O San Diego 20%
(heat input basis)
Tangential mode
1.50 x 106 BTu/hr
300°F Secondary Air
Figure 4-22. NO emissions during baseline testing (San Diego)
-------
200
OJ
o
Q.
Q.
TOO
ro
0
Data from previous work
17
18
19 20 21
Stack Gas Temperature °F x 102
22
Tangential Mode
Gas Firing
1.0 x 106 Btu/hr
20% Excess Air
O Baseline Point
Figure 4-23. Thermal NO (previous work).
-------
believed to be the result of lower thermal NO contributions resulting from
cooler flame temperatures. The 20 percent Richmond/gas flame was extremely
luminous which resulted in higher radiation losses from the flame and an
overall cooler flame. It is well documented that thermal NO is very sensi-
tive to temperature.
Note that neither the Americology nor the San Diego fuel curves
contain 5 percent by heat input refuse concentration fuel mixtures. This
is due to the density of these two fuels. The feed systems were not capa-
ble of delivering a consistent feed at the required low flowrates.
A comparative analysis of the combustion characteristics of each of
the individual fuel types is illustrated in Figure 4-24. Shown in this
figure are curves representing a constant fuel mixture consisting of na-
tural gas and each of the refuse types with all the other parameters held
constant. The fuel nitrogen content of each mixture is listed in the
legend. The order of the curves in Figure 4-24 demonstrates the fact that
each refuse contributed to the overall NO level in a unique manner. While
the curves representing the Ames and San Diego source mixtures are consis-
tent with the chemical relationship, the Richmond source fuel is clearly
varying in fuel nitrogen concentration.
As noted earlier, all sulfur dioxide data is valid only on a rela-
tive basis. However, a good comparison of fuel types is illustrated in
Figure 4-25, where curves for each refuse type cofired with gas at 30 per-
cent excess air are plotted. The sulfur analysis of each fuel is also
listed. These curves demonstrate the unique characteristics which each
refuse exhibits in a combustion environment.
125
-------
N
zoo
ro
CVJ
o
0.
Q.
100
0
20% Refuse/natural gas
I
I
I
10 20 30
Excess air (percent)
O
Q
D
Ames
Richmond
Americolorjy
San Dieqo
0.73
0.83
0.91
0.69
Tangential Mode
1.50 x 106 BTu/hr
300°F Secondary Air
Figure 4-24. NO emissions during baseline testing (all RDF's)
-------
175 r
125
CM
o
100
CL
Q-
CO
O
QJ
Tangential mode
Gas co-fire
1.5 x 10s Btu/hr
30% excess air
Ames
Richmond
Americology
San Diego
S (%)
0.35
°-16
0.17
0.21
75
OJ
ac.
50
25
5 10 15 20
Refuse concentration, percent (h.v. basis)
Figure 4-25. S02 data (all RDF's).
127
-------
All gaseous emission data for natural gas testing is listed in the
appendix for completeness. In general, however, carbon monoxide, carbon
dioxide and unburned hydrocarbon measurements were consistently low through-
out the baseline tests.
The refuse/coal cofired tests were focused on obtaining stack gas
analyses other than gaseous emissions. However, gaseous emissions were
recorded and are presented in the appendix for completeness. The NO emis-
sions for coal cofiring are summarized in Figure 4-26 where the effect of
excess air percentage and refuse concentration in the fuel mixture are
illustrated. As is shown, a general downward trend is exhibited as refuse
concentration is increased. This trend was first believed to be the result
of cooler flame temperatures reducing the thermal NO and a reduction in
the amount of fuel N available. However, an examination of the fuel nitro-
gen availability using thermal NO data taken from Figure 4-23 as a function
of temperature indicates that a reduction in fuel nitrogen conversion is
the likely source of lower NO levels. This data is shown schematically
in Figure 4-27 where fuel nitrogen conversion and fuel nitrogen availability
is plotted.
4.5.3 Particulate Analyses
The results of the particulate analyses are presented according to
variation of combustion conditions.
4.5.3.1 Refuse Type
Table 4-8 lists the results as a function of refuse type for the gas
cofired tests. Each point is also expressed as percent of total mass to
illustrate where the bulk of the loading lies, according to size. As noted
128
-------
ro
10
CM
o
E
ex
Q.
600
300
200
100
Richmond refuse/Pittsurgh coal
I
D
I
I
10 20 30
1 RDF (heat input)
D 107, Excess Air
<£> 20Z Excess Air
Q 30'^ Excess Air
Tangential Mode
1.50 x 106 BTu/m
300"F Secondary Air
Figure 4-26. NO emissions during detailed testing (Richmond
RDF/Pittsburg coal).
-------
CO
O
O>
c
O
(U
en
o
u.
r- 5.0
O)
3
2.5
Q Nitrogen Conversion
tD Nitrogen Available
Tangential mode
1.5 x 106 Btu/hr
300°F Secondary air
10% Excess air
2.0
Ot
cr
n>
c
ro
1.0 o
to
cr
I
I
10 20
Refuse content, % (heating value basis)
Figure 4-27. Fuel nitrogen contribution.
-------
TABLE 4-8. PARTICIPATE ANALYSES: EFFECT OF RDF TYPE
Type/
% RDF
Ames/
20%
Richmond/
20%
Americology/
20%
San Diego/
20%
Ames/
10%
Richmond/
10%
Americology/
10%
San Diego/
10%
Test*
No.
13A
% of Total
13B
% of Total
13C
% of Total
13D
% of Total
11A
% of Total
11B
% of Total
nc
% of Total
11D
% of Total
Excess
Air
(%)
20
Filter
(gr/ft3)
0.03912
68.96
0.03223
90.94
0.04063
85.65
0.06158
79.59
0.03165
81.51
0.02445
94.04
0.06330
87.05
0.06684
88.32
>io
(gr/ft3)
0.01067
18.81
0.00044
1.24
0.00233
4.91
0.00732
9.46
0.00317
8.16
0.00095
3.65
0.00283
3.89
0.00354
4.68
>3u
(gr/ft3)
0.00343
6.05
0.00042
1.19
0.00232
4.89
0.00245
3.17
0.00213
5.49
0.00036
1.38
0.00288
3.96
0.00137
1.81
>lu ,
(gr/ft3)
0.00351
6.19
0.00235
6.63
0.00217
4.57
0.00601
7.77
0.00169
4.35
0.00023
0.88
0.00370
5.09
0.00391
5.17
Total
(gr/ft3)
0.05673
0.03544
0.04744
0.07737
0.03883
0.02600
0.07272
0.07568
% Ash
22.38
12.25
27.55
38.05
22.38
12.25
27.55
38.05
Fired with natural gas
-------
in Section 4.2.5, the highest percentage loading consistently occurred
in the less than 10 micron (y) range which was trapped in the filters. As
can be noted from the table, the grain loading corresponds roughly to the
percent ash in the fuel. It can also be observed that the Richmond fuel
consistently had the highest percentage of particules in the less than 1 y
size cut. Similarly, the Ames fuel consistently had the lowest percentage
in this range size cut.
4.5.3.2 Excess Air
Table 4-9 shows the effect of excess air for the coal cofired tests.
Both 10 percent and 20 percent refuse concentration points are shown. Note
that in the coal cofired tests, the majority of the loadings were evenly
distributed in the size ranges larger than 1.0 micron (y). Several addi-
tional comments can be made regarding Table 4-9.
The percent material in the less than 1 y size cut increases
with excess air. If the particulate is friable, the increase
in velocity may cause more of the material to break up into the
smaller size fraction.
The total grain loadings decreased as the excess air increased,
but more rapidly than straight dilutions would account for. This
could indicate that there is more unburned carbon in the particu-
late at the lower excess air levels. Thus it appears that the
effect of excess air is to lower the overall grain loading while
concentrating more of the particulate in the respirable size
fraction.
132
-------
TABLE 4-9. PARTICULATE ANALYSES: EFFECT OF EXCESS AIR
co
CO
Test
No.
37
% of Total
38
% of Total
34
% of Total
32
% of Total
31
% of Total
Excess
Air
(«)
10
20
30
10
20
y*
h
RDF
10
20
Filter
(gr/ft3)
0.02287
2.90
0.04439
7.60
0.04833
9.25
0.05455
5.27
0.02583
9.13
>10u
(gr/ft3)
0.39145
49.59
0.26900
46.07
0.25437
48.49
0.45018
43.51
0.13386
47.34
>3y
(gr/ft3)
0.29222
37.02
0.22627
38.75
0.20814
39.82
0.39695
38.37
0.10736
37.97
>lu ,
(gr/ft3)
0.08274
10.48
0.04422
7.87
0.03725
7.13
0.13290
12.85
0.01570
5.55
Total
(gr/ft3)
i
0.78930
0.58389
0.52276
1.03459
0.28276
Fired with coal
-------
4.5.3.3 RDF Concentration
The effect of refuse concentration, cofired with coal, is shown
in Table 4-10. Results are listed for both 10 and 20 percent excess air
levels. An even distribution again occurred in the larger than 1.0 micron
(y) size ranges. This table is merely a rearrangement of the previous
table. The only point that needs to be reemphasized here is that it ap-
pears that the fraction in the less than 1 y size cut increases with in-
creasing percent RDF. However, these conclusions should be addressed with
a bit of caution because the total grain loadings did not increase with per-
cent RDF in all cases. The reason for this apparent data scatter is not
clear at this time. However, it could be caused by holdup in the heat
exchange sections of the furnace, by non-isokinetic sampling in the duct,
or by the wall and probe effects in the small exhaust duct due to the stand-
ard large EPA method 5 sampling probe.
A summary curve of the data shown in the table is shown in Figure
4-28 as the cumulative percent less than a given particle size. This again
shows the trend of a higher percentage in the less than 1 y size cut as
the percent of RDF increases.
4.5.3.4 Fuel Makeup
Table 4-11 compares the particulate loadings of the three fuel mix-
tures. These results are further illustrated in Figure 4-29 where particu-
late loading is plotted along with fuel ash content. The table illustrates
that the fraction in the less than 1 y size cut is increased when firing
coal alone. This indicates that the RDF is contributing to this fraction
and probably not agglomerating to the larger coal particles. The figure
134
-------
TABLE 4-10. PARTICULATE ANALYSES: EFFECT OF PERCENT RDF
to
Test
No.
37
% of Total
32
% of Total
38
% of Total
31
% of Total
35
Excess
Air
(%)
10
20
20
20
*
%
RDF
10
20
10
20
30
Filter
(gr/ft3)
0.02287
2.90
0.05455
5.27
0.04439
7.60
0.02583
9.13
>10u
(gr/ft3)
0.39145
49.59
0.45018
43.51
0.26900
46.07
0.13386
47.34
>3y .
(gr/ft3)
0.29222
37.02
0.39695
38.37
0.22627
38.75
0.10736
37.97
>lu ,
(gr/ft3)
0.08274
10.48
0.13290
12.85
0.04422
7.57
0.01570
5.55
Total
(gr/ft )
0.78930
1.03459
0.58389
0.28276
*Fired with coal
-------
TABLE 4-11. PARTICULATE ANALYSES: COAL VS. 10% RDF + COAL VS. 10% RDF + GAS
co
cr>
Test
No.
40
% of Total
28
% of Total
11B
% of Total
Excess
Air
(*)
20
% RDF
+ Fuel
Coal
10% RDF +
Coal
10% RDF +
Gas
Filter
(gr/ft3)
0.02052
2.14
0.04439
7.60
0.02445
94.04
>10
(gr/ft3)
0.53895
56.28
0.26900
46.07
0.00095
3.65
>3 T
(gr/ft3)
0.34355
35.87
0.22627
38.75
0.00036
1.38
>1 7
(gr/ft3)
0.05465
5.71
0.04422
7.57
0.00023
0.88
Total _
(gr/ft3)
0.97769
0.58389
0.02600
-------
100
80
60
40
20
310
8
6
u
f-
o>
O
I-
ID
1.0
0.8
0.6
0.4
0.2
0.1
20% EA
0 Test 38
^ Test 31
& Test 35
0 Test 40
'
% RDF
10
20
30
0
' _ « I ' l ' L
0.01 0.1 0.5 2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-28. Stack gas particle size vs. cumulative percent
less than diameter.
137
-------
00
_ 16.0
_ 12.0 T
10 20 30
Refuse concentration (% heat input basis)
in
_o
c
CD
8.0 g
o
c/l
fO
- 4.0
0)
3
TANGENTIAL MODE
1.50 x 106 BTU/hr
300 F Secondary Air
2Q% Excess Air
Richmond Refuse/
Pittsburgh Coal
Particulate Loading
Ash Content
Figure 4-29. Particulate loading results.
-------
indicates a rather strange effect, and that is that the grain loadings
decrease with increasing percent RDF with a minimum at 20 percent RDF.
Possible explanations for this include a greater hold up in the convective
section, or more material reaching the ashpit or sticking to the walls of
the furnace. It is possible that resultant ash properties or heat trans-
fer conditions are changing such that more material is deposited either
in the furnace or on the convective tubes. However, the duration of each
of these tests was not sufficiently long to determine if this hypothesis
is true. In addition, due to the refractory walls and dissimilar convec-
tive tubes compared to a full-scale boiler, it is rather speculative to
say a similar effect would occur in the full-scale systems.
4.5.3.5 Percent Combustibles in Flyash
Table 4-12 lists the results of the analysis on percent combustibles
as a function of percent excess air and percent refuse when cofired with
coal. As was the case for CO and unburned hydrocarbons, these results
indicate that the combustion efficiency is quite good in the pilot-scale
facility when cofiring RDF with coal as long as the excess air is above
10 percent. There is an indication that even this facility does not oper-
ate quite as efficiently with the refuse as with coal alone. This certainly
has been the case in full-scale units where considerable unburned material
has found its way to the ashpit. The reference to the plugging of the ash-
pit in Section 4.2.5 is another indication of unique problems with the RDF
materials.
However, it is possible that the additional shredding and/or the
hot refractory walls aid in ignition and achieving complete combustion in
139
-------
TABLE 4-12. COMBUSTIBLE CONTENT IN FLYASH
l/l
r-
in
TO
CO
+J
3
Q.
C
» t
4->
ro
OJ
31
Z
0
1-^
«t
a:
I
z
UJ
o
ST.
O
o
U-
D
cc:
*s
o
^s
0
&s
o
00
frS
0
ro
EXCESS AIR
10% 20% 30?;
0.43%
9.56% 1.37%
8.15% 1.35%
1.17% 3.46%
Tangential mode - 1.5 x 106 Btu/hr
Richmond refuse/Pitts #8 coal
300°F sec air
140
-------
the pilot-scale facility. Perhaps boilers designed specifically to burn
RDF cofired with coal will require a hotter radiative section. Of course,
the resulting ashing problems associated with the coal would have to be
taken into consideration.
4.5.4 Trace Metals
Concentrations of 11 trace metals were determined in the solid par-
ticulate and condensible vapors collected in the SASS impingers.
A summary of the total concentrations found on a yg/Btu basis is
shown in Table 4-13 for all the coal plus RDF tests where the SASS train was
used. A comparison is also made with the gas test using the same Richmond
RDF. Although few conclusions can be drawn with regard to this limited
sample, the following commments are in order:
With a few exceptions, the order of magnitude of each of the
trace elements does not vary greatly from test to test.
Exceptions to this comment include the following:
Cu Tests #37 and #40
Zn Test #32
Pb Test #40
Sn Test #llb
As Test #32
t There appears to be no clear trend on any of the elements with
regard either to percent RDF or percent excess air.
There does not appear to be much difference in the total trace
metal concentrations when firing gas plus RDF or coal plus RDF
(six approximately the same, one higher, and four lower).
This last comment leaves the validity of these measurements somewhat in
question as it would have been expected that the trace metal concentrations
when cofiring with natural gas would be considerably lower.
141
-------
TABLE 4-13. TOTAL TRACE METAL LOADINGS (yg/Btu) - COAL COFIRING
Fuel
% EA
% RDF
Coal
10
10
Coal
20
10
Coal
10
20
Coal
20
20
Coal
30
30
Coal
20
0
Gas
20
10
ro
^"v^est No.
Elemeru^^^
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
#37
<1.660
1.978
<0.207
0.697
<0.0081
0.004
<1.125
<0.0047
<0.1101
0.0274
<0.0389
#38
<0.2693
<0.0090
<0.0913
<0.0173
<0.0323
#32
<0.309
5.327
<0.520
1.319
0.0126
<0.0009
<1.2525
<0.0386
0.1060
<0.0293
1 . 3839
#31
<0.184
1.504
<0.3549
<2.591
<0.013
<0.0108
<0.4379
<0.0809
0.1508
<0.0058
<0.0408
#34
0.206
2.025
<0.3512
2.865
0.0037
0.0025
<1.382
<0.010
O.1481
0.0187
<0.0402
#40
1.958
0.686
0.153
17.53
0.009
<0.0176
<1.7540
<0.024
0.130
<0.0015
<0.088
#11 B
0.340
1.821
0.0209
1.500
0.0062
<0.0034
<0.0277
0.0333
<3.503
<0.0009
<0.0184
-------
However, the nonhomogeneity of the material must be considered as
well as the influence of the test furnace. First, it is possible that
large concentrations of a particular trace metal can be present locally
in the feed and find their way to the stack sampling equipment. Holdup
of material in the heat exchange sections of the furnace system can also
result in momentary high particulate concentrations if the material breaks
loose from the heat exchange surfaces in large discrete clumps. Finally,
metals in the furnace from the burners (particularly copper, lead, and Zn
from cooling coils, silver solder and brazing compounds) may also find
their way to the stack. Due to these factors, it will probably require a
large data base at any given test condition to obtain a statistically mean-
ingful result.
Table 4-14 lists the percentages of the total trace metals found as
condensible vapors collected in the organic module and impinger sections
of the SASS. Again, no clear trends are present although Hg, Cu, Mn, and
Sn generally had high percentages in the vapor phase. Cd was usually split
between the vapor and solid and As was almost always found with the par-
ticulate. The remaining trace metals had widely varying concentrations
of the condensible material.
Table 4-15 presents the total trace metal concentration for the four
RDF materials when cofired with natural gas. Again, there appears to be
wide variations between the different RDF types. Similarly, Table 4-16
presents the percent vapor for each of these materials. Hg always appears
in the vapor and As in the particulate. 8e, Cu, Sn and Mn were also usu-
ally found in the vapor. Again, the heterogeneous nature of these materials
must be considered.
143
-------
TABLE 4-14. TRACE METAL CONCENTRATIONS (%) - COAL COFIRING
Fuel
% EA
% RDF
Coal
10
10
Coal
20
10
Coal
10
20
Coal
20
20
Coal
30
30
Coal
20
0
Gas
20
10
^\[est No.
Element^v^^
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
#37
99.9
81.1
90.6
57.4
43.2
90.0
1.1
17.0
74.8
31.0
3.9
#38
91.7
1.1
63.9
91.3
10.2
#32
79.1
28.2
89.5
28.1
51.6
11.1
0.9
76.9
68.5
97.6
0.3
#31
45.4
44.9
92.3
3.8
2.3
41.7
30.0
61.8
49.5
100
3.7
#34
73.1
68.2
93.2
84.5
49.9
23.0
2.0
28.7
87.6
81.3
9.3
#40
96.2
77.2
39.4
98
48.4
73.9
25.1
3.3
65
80
18.2
#11B
64.7
24.5
77
40.6
43.5
82
64.1
7.2
99.6
100
9.2
-------
TABLE 4-15. TOTAL TRACE METAL LOADINGS (yg/Btu) - GAS COFIRING
Fuel
%EA
% RDF
Gas
20
10
Gas
20
10
Gas
20
10
Gas
20
10
^^^v^est No.
Elementr^v.
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
11A
Ames
0.653
2.972
o.om
4.38
0.0119
0.0008
0.088
0.118
0.090
0.032
0.020
11B
Richmond
0.340
1.821
0.021
1.50
0.006
0.003
0.028
0.033
3.50
0.001
0.018
nc
Americology
0.039
0.222
0.102
0.005
0.0095
11D
San Diego
4.11
13.58
0.238
42.08
0.0613
0.00027
0.281
0.0988
0.722
0.035
0.319
145
-------
TABLE 4-16. TRACE METAL CONCENTRATIONS AS VAPOR (%) - COAL COFIRING
Fuel
% EA
% RDF
Gas
20
10
Gas
20
10
Gas
20
10
Gas
20
10
""^^Jest No.
Elemeiu^v^^
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
11A
Ames
80.2
21.7
78.4
7.5
5.9
94.9
11.3
0.6
68.2
99.9
11.1
11B
Richmond
64.7
24.5
77
40.6
43.5
82
64.1
7.2
99.6
100
9.2
nc
Americology
48
1.2
89
98.8
30.2
nc
San Diego
86.8
8.7
92.3
10.7
98.2
94.8
6.4
39.7
15.7
99.9
55.8
146
-------
A comparison was also made between the trace metal concentrations in
the participate flyash found in these tests and data found in the literature
for both coal and coal plus various RDF. These results are shown in Table
4-17 for each of the test conditions and for three sets of field data. The
field data is from the St. Louis demonstration (Reference 12), Wright Patter-
son Air Force Base (Reference 13), and the Ames-Iowa facility (Reference 14).
The data is presented in yg/grain of flyash for the solid particulate only.
Again, wide variations in both the data developed on this program as well
as the field data are seen. It should also be mentioned that the field
data are average numbers and that there was considerable variation even
from one site. From the field data, it appears the results generated here
are within the same order of magnitude. However, trends as a function of
either excess air or percent RDF still cannot be discerned.
Finally, two sets of particulate data were analyzed for each of the
trace metals in each of the cyclone size cuts. This was done for the coal
only test (#40) and for coal plus 20 percent RDF at 20 percent excess air,
Test 31. Figures 4-30 through 4-40 show the charts of cumulative percent
versus size cut for each trace metal. As before, this is plotted as the
cumulative percent below and including a given size. The first cyclone
catches all material >10 y, and the filter catches everything less than 1 y.
Seven out of the 11 elements indicate that the presence of RDF results in
a higher percentage in the smaller size cuts. Trace metals which have a
reverse trend include As, Be, Mn, and Zn where the coal only has high con-
centrations of these elements in the finer sizes. However, in light of the
randomness of much of the other trace metal data, caution should be exercised
in drawing any definitive conclusions from these curves.
147
-------
TABLE 4-17. TRACE METAL CONCENTRATIONS (yg/g OF FLYASH) -PILOT VS. FULL SCALE (PARTICULATE ONLY)
Fuel Coal
% EA 10
% RDF 10
Coal
20
10
Coal
10
20
Coal
20
20
Coal
30
30
Coal
20
0
Coal
7
Coal
0
Coal
34
Coal
51
Coal
0
Coal
20
Coal
50
Coal
0
Test
Element
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
#37
1,508
506.8
26
402
6.2
0.6
1,508
5.31
<38
25.5
50.6
#38
--
76.5
--
--
--
30.3
113
5.1
99
#32
4,799
69
1,190
7.6
1.0
1,558
11.1
<42
8.5
55.9
#31
4,611
28,927
619
118,475
515.4
83.8
12,188
1415.7
4,042
1.4
1206.6
#34
126
1,465
54
1,010
4.1
4.3
3,081
16.2
4.2
8.0
82.9
#40
1,304
2,516
600
5,519
42.9
50.0
4,994
29.8
596
1.1
725.7
St. Louis
Ref 12
430
2,534
--
1,681
44
24.3
12,050
17.3
--
13.04
62
St. Louis
Ref 12
236
1,102
--
598
35
8.98
2,584
1.82
6.42
189
WPAFB
Ref 13
--
11,433
--
9,880
--
--
--
--
WPAFB
Ref 1"+
--
27,563
21,290
--
--
--
--
-
WPAFB
Ref 13
--
902
--
493
--
--
Ames
Ref l1*
472
29,211
414
31 ,684
--
--
3,196
--
--
--
Ames
Ref !"«
379
25,211
360
22,815
--
--
235,710
--
--
~~
Ames
Ref Vt
153
8,373
628
6,733
--
3,625
--
--
--
oo
-------
TOO
80
60
40 h
20
10
8
Ol
fM
0)
(_>
1.0
0.8
0.6
0.4
0.2
0.1
- - 20:: RDF + Coal
Coal
TRACE METAL - Cu
Ml II I 1 1 i i I I i i i
J L
I I
JL_L
0.01 0.1 0.5 2
10
30 50 70
90
98
S9.fi 99.99
Cumulative percent less than diameter
Figure 4-30. Stack gas particle size vs. cumulative percent
less than diameter -trace metal Cu.
US
-------
£
OJ
tM
to
o>
O
4->
s_
no
Q.
100
80
60
40
20
10
8
6
4
1
0.8
0.6
0.4
0.2 -
20% RDF + Coal
Coal
TRACE METAL - ZN
o.i
0.01 0.1 0.5
i I i I i l i
j i i i
2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-31. Stack gas particle size vs. cumulative percent
less than diameter -trace metal Zn.
150
-------
100
80 -
60 -
40 -
20
10
8
c
o
t.
o
o;
ISI
Ol
o
0.8
0.6
0.4
0.
0.
20% RDF + Coal
Coal
TRACE METAL - MN
-L-LJ' i ' ' ' i I I i i I
' ' ' '
11
0.01 0.1 0.5 2
10
30 50 70
90
98
99.9 99.99
Cumulative percent less than diameter
Figure 4-32. Stack gas particle size vs. cumulative percent
less than diameter trace metal Mn.
151
-------
to
c
o
S-
u
QJ
NJ
OJ
O
4-J
s-
o.
100
80
60
40
20
10
8
6
1.0
0.8
0.6
0.4
0.2 -
20% RDF
Coal
coal
TRACE METAL - PB
0.1
1 1 1
1 1 1
II
0.01 0.1 0.5 2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-33. Stack gas particle size vs. cumulative percent
less than diameter trace metal Pb.
152
-------
to
c
o
o
-------
100
80
60
40
20
^ 10
~ 8
6
to
c
o
s-
u
^
E
c
o
re
O-
1.0
0.8
0.6
0.4
0.2
0.1
20% RDF + coal
Coal
TRACE METAL - BE
0.01 0.1 0.5 2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-35. Stack gas particle size vs. cumulative percent
less than diameter - trace metal Be.
154
-------
100
80
60
40
20
3: 10
0)
N
I/)
s-
re
1.0
0.8
0.6
0.4
0.2
0.1
- - 20% RDF + coal
Coal
TRACE METAL - Ti
0.01 0.1 0.5 2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-36. Stack gas particle size vs. cumulative percent
less than diameter trace metal Ti.
155
-------
100
80
60
40
20
10
~ 8
co
§ 6
1 4
-------
100
80
60 Coal
40
20
~ 10
8
o c
j- b
'i
0)
N
1 2
£ l.C
0.
0.6
0.
0.
0.
- - 20% RDF + coal TRACE METAL - SN
0.01 0.1 0.5 2 10 30 50 70 90 98 99988 99.99
Cumulative percent less than diameter
Figure 4-38. Stack gas particle size vs. cumulative percent
less than diameter - trace metal Sn.
157
-------
100
80
60
40
20
3 10
8
o
N
(/)
O)
u
J-
1C
a.
1.0
0.8
0.6
0.4
0.2
0.1
- - 20% RDF + coal
Coal
TRACE METAL - HG
0.01 0.1 0.5 2 10 30 50 70 90 98
Cumulative percent less than diameter
99.8 99.99
Figure 4-39. Stack gas particle size vs. cumulative percent less
than diameter - trace metal Hg.
158
-------
100
80
60
40
20
10
^ 8
CO r
C 0
O.
rs,
^ 2
QJ
U
to
Q.
1.0
0.8
0.6
0.4
0.2
0.1
20% RDF + Coal
Coal
iii
TRACE METAL - As
0.01 0.1 0.5 2
10
30 50 70
90
98
Cumulative percent less than diameter
99.8 99.99
Figure 4-40. Stack gas particle size vs. cumulative percent
less than diameter - trace metal As.
159
-------
In summary, it appears that very little can be said about this data
with regard to either the levels or trends of trace metals when cofiring
RDF with either coal or natural gas. For future tests, it is recommended
that at least five samples be collected at any given test condition in order
to adequately determine the concentrations. In addition, background tests
on gas only also need to be taken so that metals coming off the furnace
can be taken into account.
4.5.5 Organics
As was mentioned in Section 4.4, the organic modules of the SASS
train were analyzed by GC/MS for organic compounds. Tests 31, 32, 34, 37,
and 40 contained no detectable organic compounds. Samples from Tests 38,
11A, 11B, and 11C contained polynuclear aromatic hydrocarbons and deriva-
tives in the amounts indicated in Table 4-18. No PCBs were detected in
any samples.
Two other compounds were detected in the RDF Test 11B sample. The
mass spectra of these components were indicative of silicon containing com-
pounds. They could not, however, be positively identified. The spectra
of these compounds as well as the total ion current traces for the analyses
are available if needed.
A final point involves the presence of medium weight polynuclear
aromatic hydrocarbons in these stack samples. A large volume of literature
indicates that combustion of hydrocarbon fuels gives rise to polynuclear
aromatic hydrocarbons and also to highly polymerized species which are
collectively known as "soot." The latter species are not readily analyzed,
but the lower homologues are analyzed as the polynuclear aromatics. In
these samples, the medium weight species such as pyrene, fluoranthene and
160
-------
TABLE 4-18. ORGANICS FOUND
Test Condition
Organic
Amount
Gas Cofire
10% RDF
20% EA
Ames Fuel
Gas Cofire
10% RDF
20% EA
Richmond Fuel
Gas Cofire
10% RDF
20% EA
Americology Fuel
Coal Cofire
10% RDF
20% EA
fluoranthene
pyrene
phenanthrene
fluoranthene
pyrene
diphenyl ether
biphenyl phenyl ether
phenanthrene
pyrene
phenanthrene
0.0000102 pg/Btu
0.0003325 ug/Btu
0.0000641 yg/Btu
0.0001601 vg/Btu
0.0005765 yg/Btu
0.003395 yg/Btu
0.001697 yg/Btu
0.0000593 yg/Btu
0.0010369 yg/Btu
0.0000981 yg/Btu
161
-------
phenanthrene normally dominate with the higher molecular weight species
(such as benzo(a) pyrene present also, but at concentrations lower by a
factor of 10 to 100. If such were true with the RDF samples, then these
carcinogenic compounds would be present, but at concentrations below the
detection limit for these analyses.
In addition, it should be remembered that only two of the LC frac-
tions were analyzed (LC 2,3). Tables 4-19 and 4-20 show the quantity and
percent of the material found in all of the LC fractions for each of the
tests where a SASS analysis was made. As can be seen from this table, con-
siderable material was found in Fraction LC 1, 6 and 7 in many of the tests
although these fractions were not analyzed. Table 4-21 gives a representa-
tive listing of the possible compounds that could make up each of these
fractions and the MEG concentration limit. Thus, if the material in these
fractions were made up of any one of these compounds, it could exceed the
MEG criteria. For this reason alone, further analysis on these samples
is warranted.
162
-------
TABLE 4-19. LC COLUMN DATA
CO
Test
No.
HA
11B
TIC
31
32
34
37
38
40
mg/m3
Ll
0.10728
0.49074
0.16019
0.12811
0.91087
0.49676
0. 38268
1.12582
0.21218
L2
0.01314
0.05708
0.03620
0.00217
0.04315
0
0.00656
0.02124
0.05378
L3
0.08393
0.24537
0.08688
0.03908
0.12758
0
0.02697
0.07379
0.06410
L4
0.09414
0.15677
0.05249
0.03474
0.15572
0.04909
0.07070
0.13416
0.08767
L5
0.08612
0.26326
0.10951
0.03908
0.15197
0.07034
0.09767
0.12522
0.14146
Lfi
0.05692
1.36317
0.05068
0.01954
0.17448
0.02125
0.04155
0.09503
0.90547
L7
1.14943
0.84943
1.11047
0.38144
1.77577
1.20453
1.25519
2.46853
3.79500
-------
TABLE 4-20. LC COLUMN DATA
CT>
-p.
Test
No.
11A
11B
11C
31
32
34
37
38
40
LI/LT
%
6.74
14.32
9.97
19.89
27.28
26.97
20.34
27.84
4.03
4/4
%
0.83
1.67
2.25
0.34
1.29
0
0.35
0.53
1.02
L3/LT
%
5.27
7.16
5.41
6.07
3.82
0
1.43
1.82
1.22
VLT
%
5.92
4.58
3.27
5.39
4.66
2.67
3.76
3.32
1.67
L5/LT
%
5.41
7.68
6.82
6.07
4.55
3.82
5.19
3.10
2.69
VLT
%
3.58
39.79
3.15
3.03
5.22
1.15
2.21
2.35
17.22
L?/LT
%
72.25
24.79
69.13
59.21
53.17
65.39
66.72
61.05
72.15
-------
TABLE 4-21. POSSIBLE COMPOUNDS IN LC FRACTIONS NOT ANALYZED
Test
No.
38
40
ne
Sample
Fraction
LCI
LC7
LCI
LC6
LC7
LCI
LC6
LC7
Concentration
(pg/m3)
1125.82
2468.53
212.18
905.47
3795.00
490.74
1363.17
849.32
Sample Fraction
Tetraethyllead
2 ,4 ,6-Tri ni trophenol
4,6-Dinitro-O-Cresol
4,4'-Methylene-Bis-(2-Ch1oroaniline)
Penthachlorophenol
1 -Ami nonaphtha 1 ene
Dinitro-P-Cresol
Dini trophenol s
Tetraethyllead
2- Ami nonaphtha 1 ene
Dibenz (A,H) Acridine
Dibenz (A,J) Acridine
Anisidines
Perchloromethanethiol
2,4,6-Trinitrophenol
4,6-Dinitro-O-Cresol
4,4'-Methylene-Bis-(2-Chloroaniline)
Penthachl orophenol
1-Aminonaphthalene
Dinitro-P-Cresol
Din1 trophenol s
Tetraethyllead
2-Aminonaphthalene
Dibenz (A,H) Acridine
Dibenz (A,J) Acridine
Anisidines
Perchloromethanethiol
Dibenzo (C,D) Carbazole
Methyl arolne
2,4,6-Trinitrophenol
4,6-Dinitro-O-Cresol
4,4' -Methylene-Bis-(2-Chloroani line)
Penthachl orophenol
1 -Ami nonaph thai ene
Dinltro-P-Cresol
Concentration
Limit
(ug/m3)
100.0
100.0
200.0
220.0
500.0
560.0
680.0
1400.0
100.0
170.0
220.0
250.0
500.00
800.0
100.0
200.0
220.0
500.0
560.0
680.0
1400.0
100.0
170.0
220.0
250.0
500.0
800.0
1000.0
1200.0
100.0
200.0
220.0
500.0
560.0
680.0
165
-------
REFERENCES
1. Brown R. A., Kelly, J. T., Neubauer, Peter, "Pilot Scale Evaluation of
NOX Combustion Control for Pulverized Coal, Phase II Final Report."
EPA 600/7-79-132, June 1979.
2. Wendt, J. 0. L., Lee, S. W., Pershing D. W., "Pollutant Control Through
Staged Combustion of Pulverized Coal. Phase I -- Comprehensive Report.
U.S. Dept. of Energy, Fe-1817-4, February 1978.
3. Johnson, S. A., Cioffi, P. L., McElroy, M. W., "Development of an Ad-
vanced Combustion System to Minimize NOX Emissions from Coal-Fired
Boilers." Presented to 1978 Joint Power Conference, Dallas, Texas,
September 11, 1978.
4. Demeter, J. J., et al., "Combustion of Coal-Oil Slurry in a 100-HP
Firetube Boiler," PERC/R1-77/8, Pittsburgh Energy Research Center,
Pittsburgh, Pennsylvania, May 1977, pp. 3-8.
5. BeeV, J. M., Combustion Aerodynamics, John Wiley and Sons, New York,
N.Y., 1972.
6. Thompson, R. E., et al., "Effectiveness of Gas Recirculation and Staged
Combustion of Reducing NOX on a 560-MW Coal-Fired Boiler," EPRI FP-257,
September 1976.
7. Heap, M. P., et al., "The Optimization of Burner Design Parameters to
Control NOX Formation in Pulverized Coal and Heavy Oil Flames," Pro-
ceedings of the Stationary Source Combustion Sumposium, Volume I, EPA-
600/2-76-1526, June 1976.
8. England, G. C., et al., "The Control of Pollutant Formation in Fuel Oil
Flames -- The Influence of Oil Properties and Spray Characteristics,"
Proceedings of the Third Stationary Source Combustion Symposium; Volume
II. Advanced Processes and Special Topics, EPA-600/7-79-0506, February
1979, pp. 41-71.
9. Brown, R. A., "Pilot Scale Investigation of Combustion Modification
Techniques for NOX Control in Industrial and Utility Boilers," EPA-
600/2-76-1526, Proceedings of the Stationary Source Combustion Sympo-
sium, Volume II, June 1976.
10. Wendt, J. 0. L. and Ekmann, J. M., "Effect of Sulfur on NOX -- Emissions
from Premixed Flames," EPA-600/2-75-075, October 1975.
11. Wendt, J. 0. L., et al., "Interactions Between Sulfur Oxides and Nitro-
gen Oxides in Combustion Processes," Proceedings of the Second Stationary
Source Combustion Symposium, Vol. IV, EPA-600/7-77-073d, July 1977.
166
-------
12. Gorman, et al., St. Louis Demonstration Project Final Report: "Power
Plant Equipment, Facilities and Environmental Evaluations," EPA Con-
tract 68-02-1871, Prepared for U.S. Environmental Protection Agency,
Washington, D.C., by Midwest Research Institute, Kansas City, Missouri,
July 1977, pp. 402.
13. Jackson, J. W., "A Bioenvironmental Study of Emissions from Refuse
Derived Fuels," USAF Environmental Health Laboratory, McClellan, Cali-
fornia, January 1976, pp. 113.
14. Hall, 0. L., et al., "Evaluation of the Ames Solid Waste Resources --
An Energy Recovery System, Part III -- Environmental Evaluation of the
Stoker-Fired Steam Generators at the City of Ames, Iowa, Prepared for
U.S. Environmental Protection Agency, Cincinnati, Ohio, and Energy
Research and Development Administration by Iowa State University, Mid-
west Research Institute, and Ames Laboratory, April 1977, pp. 133.
167
-------
APPENDIX
DATA SUMMARY - DISTRIBUTED AIR
DATA SUMMARY - COAL/OIL MIXTURE
DATA SUMMARY - RDF TESTING
COM/DOE REPORT
169
-------
TABLE A-l. DATA SUMMARY - DISTRIBUTED AIR
Test
No.
209a
b
c
d
e
f
g
h
i
j
k
i
21 Oa
b
c
d
e
f
g
h
211a
b
c
d
e
f
g
h
1
j
k
n
212a
b
c
d
e
f
g
h
i
Fuel
1
SR
0.80
0.80
0.80
0.80
0.80
0.80
0.80
0.95
0.95
0.80
0.80
0.80
0.95
0.95
0.95
0.80
0.80
0.80
0.80
0.95
0.80
0.80
0.80
0.80
0.80
0.80
0.95
0.95
0.95
0.80
0.80
0.80
0.80
0.80
0.80
0.95
0.95
0.95
0.95
0.95
0.95
EA
%
15
Load
xlO Btu/hr
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
0.85
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
0.85
0.85
0.85
0.85
0.85
0.85
1.3
1.3
1.3
Preheat
sec °F
75
78
79
79
79
79
79
80
600
580
580
550
600
600
600
600
600
600
600
600
600
600
600
600
600
600
580
600
550
600
580
600
500
550
550
600
575
575
635
625
650
Stg °F
75
81
82
82
82
82
82
90
-_
227
269
276
350
350
350
296
296
296
322
327
323
371
381
337
336
335
288
242
230
368
422
443
256
379
409
359
334
326
330
312
305
Burners
4 IFRF
j
SW/Int
or
Yaw
4
i
i
Prim.
Stoich.
I
12
Stg Air
Mixing,'
Location
Hor b
Hor b
Hor b
Hor a
Hor a
Hor a
Hor a
Hor a
Hor a
Hor a
Hor a
Hor a
HE-J
HE-J
HE-J
HE-J
HE-J
HE-J
HE-J
HE-J
HE-K
HE-K
HE-K
HE-K
HE-K
HE-K
HE-K
HE-K
HE-K
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE-H
HE--H
Temperature
T23°F
1910
2024
2048
2077
2104
2135
2149
2165
2227
2234
2185
2176
2146
2034
2108
--
--
--
--
2362
2314
2394
2366
2374
2282
2261
2350
2325
2306
2256
2364
1877
1767
2429
2339
2621
2482
2621
2603
2577
T24"F
1957
2010
2023
2014
2023
2059
2088
2088
2088
2103
2133
2128
2132
2562
2574
2480
2508
2557
2488
2497
2540
2529
2442
2338
2368
2473
2558
2548
2438
2384
2615
2582
1927
1978
1981
1966
2003
2036
2203
2266
2268
NOC
ppm
300
310
377
299
290
257
251
424
480
262
266
250
376
451
284
133
152
163
170
258
215
260
197
185
167
224
305
252
335
289
264
229
190
219
178
222
253
296
346
289
260
Comments
SRla
0.45
0.30
0.60
0.60
0.45
0.30
0.30
0.45
0.45
0.30
0.45
0.60
0.45
0.30
0.60
0.60
0.45
0.30
0.80
0.95
0.45
0.30
0.60
0.45
0.60
0.30
0.45
0.60
0.30
0.30
0.45
0.60
0.45
0.30
0.60
0.60
0.45
0.30
0.30
0.45
0.60
1
-------
TABLE A-l. DATA SUMMARY -DISTRIBUTED AIR (CONCLUDED)
Test
No.
212j
k
t
21 3a
b
c
d
e
f
9
h
i
j
k
i
m
n
0
P
q
r
s
t
u
V
w
X
21 4a
b
c
d
e
f
9
h
Fuel
1
SR
0.95
0.95
0.95
0.80
0.80
0.80
0.95
0.95
0.95
0.95
0.95
0.95
0.80
0.80
0.80
0.80
0.80
0.80
0.95
0.95
0.95
0.80
0.80
0.80
0.95
0.95
0.95
0.95
0.95
0.80
0.80
0.80
0.80
0.95
0.95
EA
%
15
Load
« I^Btu/hr
1.7
1.7
1.7
0.85
0.85
0.85
0.85
0.85
0.85
1.3
1.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.3
.7
.7
.7
.7
.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
1.7
Preheat
sec °F
675
675
675
100
98
90
90
90
90
575
600
600
600
600
600
620
600
600
600
600
600
600
625
625
620
580
600
620
620
670
Stg °F
263
269
270
98
98
103
104
104
196
250
275
282
276
277
302
300
366
382
385
264
244
294
313
100
335
310
314
Burners
4 IFRF
SW/Int
or
Yaw
4
Prim.
Stoich.
%
12
Stg Air
Mixing/
Location
HE-K
HE-K
HE-K
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-E
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-F
HE-M
HE-M
HE-M
HE-M
HE-N
HE-N
HE-N
HE-N
Temperature
T23 °F
2878
2759
2274
2059
2089
2137
2161
2172
2188
2075
2381
1937
2017
1864
1507
2233
1760
1890
1886
2507
2886
2572
2220
2691
2855
3861
2532
T °F
>24 >-
2182
2396
2515
1830
1880
1905
1957
1974
1987
2296
2145
2258
2379
2372
2376
2379
2433
2370
2432
2456
2427
2460
2482
2597
2592
2572
2655
1996
2085
2335
2083
2235
2566
2596
2330
NOC
ppm
379
375
480
190
185
215
303
287
284
305
300
466
324
305
281
255
244
285
348
300
312
346
311
397
368
405
381
223
230
151
163
116
122
201
257
Comments
SR1a
0.60
0.45
0.30
0.45
0.60
0.30
0.30
0.45
0.60
0.60
0.45
0.30
0.30
0.45
0.60
0.60
0.45
0.30
0.30
0.45
0.60
0.45
0.60
0.30
0.45
0.60
0.30
0.60
0.70
0.60
0.70
0.70
0.60
0.60
0.70
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE
Test
No.
215a
b
c
d
21 6a
b
c
d
e
f
9
h
i
21 7a
b
c
d
e
f
g
h
i
j
k
i
m
n
0
21 8a
b
c
d
e
f
g
h
i
a
b
Residence
Time
S
Firing Rate
(Btu/hr x 106)
1.8
Radiant
Heat
Transfer
(Btu/hr x 106)
0.558
0.483
0.571
0.442
0.521
0.429
9.438
0.471
0.438
0.438
0.454
0.492
0.463
0.250
0.238
0.242
0.254
0.254
0.254
0.254
0.254
0.254
0.275
0.304
0.304
0.313
0.313
0.283
0.263
0.263
0.313
0.300
0.296
0.308
0.263
0.321
0.342
0.267
0.258
Excess
Air
20
30
40
20
20
30
20
40
40
30
20
20
40
30
20
40
40
30
20
20
20
20
20
20
20
20
20
20
30
40
20
30
40
20
20
20
20
20
20
i
Coal
Type
(X)
0
30% W.Kty.
30% Va.
30% W.Kty.
Oil
Type
Chevron
Penn
Chevron
Penn
Chevron
Penn
Preheat
Temp
80
82
77
76
82
83
82
83
82
82
81
83
83
300
Staged
Air
Preheat
Temp
277
277
277
290
216
220
Nozzle
Type
DeLavan
Sonicore
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
CO
Test
No.
219c
d
220e
f
9
h
i
j
k
i
221a
b
c
d
e
f
g
h
i
j
k
l
m
n
0
P
q
r
222a
b
c
d
e
f
g
223a
b
c
d
Residence
Time
S
L
S
L
S
L
L
S
L
S
S
Firing Rate
(Btu/hr x 106)
1.8
1.7
1.7
1.8
Radiant
Heat
Transfer
(Btu/hr x 106)
0.258
0.258
0.304
0.304
0.267
0.267
0.271
0.263
0.263
0.263
0.288
0.292
0.292
0.292
0.292
0.292
0.292
0.292
0.321
0.338
0.338
0.267
0.317
0.279
0.275
0.275
0.275
0.267
0.267
0.313
0.304
0.304
0.296
0.300
0.275
0.292
0.292
0.292
0.292
Excess
Air
20
20
20
--
20
20
20
20
20
20
30
40
20
20
20
20
20
20
20
20
20
20
20
20
20
20
30
40
20
20
20
20
20
30
40
20
20
20
20
Coal
Type
(%)
30% W.Kty.
30% Va.
30% Mont.
30% Va.
30% Mont.
0
0
0
30% Mont
Oil
Type
Penn
__'.
Chevron
Penn
Chevron
Penn
Chevron
Preheat
Temp
300
_._
300
Staged
Air
Preheat
Temp
225
228
251
246
200
238
238 -
261
261
267
274
275
275
205
205
205
220
220
261
Nozzle
Type
Sonicore
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
Test
No.
223e
f
g
h
i
j
k
i
m
n
0
P
q
r
224a
b
c
d
e
f
9
h
i
j
k
«,
225a
b
c
d
226a
b
c
d
e
f
Residence
Time
L
S
L
S
L
S
L
S
L
S
L
S
Firing Rate
(Btu/hr x 106)
1.8
1.2
1.8
Radiant
Heat
Transfer
(Btu/hr x 106)
0.292
0.292
0.292
0.292
0.292
0.292
0.292
0.292
0.292
0.292
0.288
0.288
0.288
0.288
0.304
0.304
0.308
0.304
0.304
0.304
0.304
0.304
0.308
0.308
0.283
0.288
0.313
0.313
0.313
0.313
0.313
0.317
0.300
0.300
0.288
0.288
Excess
Air
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
15
20
30
40
20
30
40
20
20
30
Coal
Type
(%)
30% Mont.
100% W.Kty.
1002 Mont.
100% Va.
1
Oil
Type
Chevron
Preheat
Temp
300
276
276
320
320
300
350
330
330
330
320
Staged
Air
Preheat
Temp
261
213
223
223
229
229
229
229
224
221
250
250
268
268
282
282
Nozzle
Type
Sonicore
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
Test
No.
21 5a
b
c
d
216a
b
c
d
e
f
g
h
i
21 7a
b
c
d
e
f
g
h
i
j
k
J.
m
n
0
218a
b
c
d
e
f
g
h
i
21 9a
b
Burner
Swirl
5
4.5
4
6.5
1.0
Atomized
Air
Flow
240
72
81
21
11
2
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
CTl
Test
No.
219c
d
220e
f
9
h
i
j
k
i
221 a
b
c
d
e
f
9
h
i
j
k
i
m
n
0
P
q
r
222a
b
c
d
e
f
g
223a
b
c
d
Burner
Swirl
1.0
1.0
0.5
0.5
Atomized
Air
Flow
166
166
150
143
1!
1
5
0
1
150
1
0
150
Atomized
Air
Pressure
22
22
26
--
22
17
14
12
22
10
22
Fuel
Pressure
26
26
27
-_
24
27
29
29
28
26
38
38
38
38
38
38
38
38
27
27
27
27
27
27
26
26
25
25
38
38
38
38
20
20
20
38
38
38
38
Fuel
Temp
200
200
210
215
210
200
210
210
200
200
200
200
200
200
200
200
200
190
200
200
200
210
220
210
210
210
200
200
200
200
200
180
180
190
200
200
200
200
Stoich.
Ratio
(SR,)
0.95
0.95
0.85
0.75
0.65
0.55
0.95
0.95
0.85
0.85
0.75
0.75
0.95
0.95
0.75
0.95
0.95
0.95
0.65
0.85
0.85
0.65
Stoich.
Ratio
(SRla)
0.65
0.55
0.55
1.20
1.20
1.20
1.20
0.55
0.65
._-
0.85
0.75
0.65
Dopant
Type
Thiophene
Thiophene
Pyridene
Pyridene
--.-
~ ~ ""
Total Fuel
Nitro/Sulfur
(*)
2.151 S
2.148 S
0.965 N
0.796 N
...
-------
TABLE A-2. DATA SUMMARY -COAL/OIL MIXTURE (CONTINUED)
Test
No.
Z23e
f
9
h
i
j
k
t,
m
n
0
P
q
r
224a
b
c
d
e
f
9
h
i
j
k
l
225a
b
c
d
226a
b
c
d
e
f
Burner
Swirl
0.5
4.0
Atomized
Air
Flow
150
110
120
110
120
190
Atomized
Air
Pressure
22
12
20
12
16
18
8.4
Fuel
Pressure
38
38
38
38
30
30
30
30
30
30
12
12
22
22
22
22
22
22
22
22
22
22
22
22
25
24
..
--
--
--
--
--
--
--
Fuel
Temp
200
200
200
200
200
200
200
200
200
200
200
200
190
190
190
190
190
180
180
180
180
180
180
180
190
180
Stoich.
Ratio
(SRj)
0.65
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.95
0.85
0.75
0.65
- --
Stoich.
Ratio
...
0.85
0.75
0.65
0.55
0.55
0.65
0.65
0.75
0.75
0.85
0.85
--.
Dopant
Type
Total Fuel
Nitro/Sulfur
(*)
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
co
Test
No.
Z15a
b
c
d
216a
b
c
d
e
f
g
h
i
217a
b
c
d
e
f
g
h
i
J
k
1
m
n
0
218a
b
c
d
e
f
g
h
i
21 9a
b
T25
2137
2049
1962
2099
1779
1732
1730
1706
1682
1790
1819
1823
1805
1816
1831
1730
1774
1899
1864
2042
2059
1834
1800
1775
1816
1865
1900
1959
1857
1856
1824
1928
2025
2061
2209
2085
1882
T26
...
1723
1811
1651
1675
1605
1529
1546
1552
1595
1611
1535
1595
1644
1580
1676
1637
1635
1602
1532
1626
1531
1663
1610
1609
1637
1578
1685
1639
1748
1732
1777
1758
1680
1678
1620
1584
1529
T27
788
834
775
766
820
763
823
793
786
787
826
823
801
846
872
815
801
776
722
770
706
907
835
839
845
793
880
886
899
930
967
862
822
825
786
784
767
°2
3.6
5.0
6.3
5.0
5.2
5.9
5.1
3.6
3.7
6.1
5.2
3.7
6.4
5.8
4.9
3.8
3.0
3.4
3.2
3.4
4.0
3.4
3.9
3.5
5.4
6.6
3.6
4.9
6.1
3.4
3.9
3.6
3.4
4.3
4.1
CO
8.5
14.5
17.5
4.7
6.8
5.8
7.6
41.7
12.6
4.8
16.0
45.6
78.0
77.1
85.0
70.0
70.2
38.2
69.5
62.5
63.8
91.5
125.7
119.8
123.0
191.0
271.0
139.0
142.0
95.0
91.0
70.7
70.4
co2
13.2
15.9
12.5
12.8
13.3
13.9
14.7
14.3
15.1
15.3
14.9
15.6
15.2
14.7
14.8
11.8
10.8
13.7
13.2
12.2
14.4
14.1
14.5
14.6
12.8
12.9
NO
649
831
890
833
451
510
460
434
358
345
333
409
487
622
508
683
385
427
380
319
328
324
265
372
364
331
332
204
760
812
688
732
763
627
571
510
416
468
422
SO
542
562
574
564
600
605
608
713
677
617
620
607
568
610
661
1756
1917
1931
1991
2017
2088
2098
2063
1963
1987
2033
691
843
548
739
786
790
780
786
790
1725
1628
UHC
2.6
0.4
0.0
1
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONTINUED)
Test
No.
21 9c
d
220e
f
9
h
i
j
k
1
221a
b
c
d
e
f
g
h
i
j
k
1
m
n
o
p
q
r
222a
b
c
d
e
f
g
223a
b
c
d
e
f
T25
2014
2025
2021
1871
1960
1898
2040
1799
1850
1835
1918
1996
2001
1917
__._
1887
1929
1977
1995
1883
1806
1834
1847
1786
1680
1947
1824
1780
1730
1922
1964
Z096
1917
1925
1972
T26
1603
7626
1637
1606
1560
1553
1691
1626
1627
1674
1602
1625
1524
1648
1615
1615
1620
1599
1632
1666
1695
1684
1639
1497
1588
1669
1648
1637
1668
1677
1601
1637
1531
1631
T27
791
803
831
831
806
817
832
885
932
909
836
832
776
835
804
790
796
781
834
850
862
862
855
836
807
867
879
915
872
815
806
790
749
823
°2
3.8
3.2
3.3
2.8
4.0
3.6
3.6
3.6
4.9
6.3
3.8
3.3
3.0
3.5
3.3
4.1
4.7
3.4
4.0
5.9
6.0
3.9
5.4
3.2
5.9
9.3
3.3
3.9
3.6
5.2
6.2
4.2
3.2
3.3
3.3
3.7
3.9
CO
64.8
60.0
47.0
75.0
73.0
66.0
83.0
113.4
99.9
83.5
82.3
166.5
105.3
132.2
1869.0
63.0
54.0
53.0
91.0
107.0
85.0
106.0
79.0
131.0
268.0
39.2
98.5
76.9
95.0
160.0
179.6
356.7
110.3
138.6
1147.0
125.7
co2
14.0
14.3
14.0
12.2
11.6
12.4
12.4
12.2
12.3
12.7
14.6
15.0
15.1
14.6
15.0
14.2
14.0
14.9
14.6
13.7
13.0
14.7
13.9
15.2
13.2
11.0
14.3
13.9
13.3
12.1
11.5
13.9
14.7
14.7
14.8
14.6
14.6
NO
404
346
389
653
623
504
413
606
705
782
622
668
580
233
600
228
606
581
541
563
607
397
622
603
471
653
533
307
226
370
399
427
630
285
566
418
280
588
SO
1554
1525
1315
843
921
913
838
538
643
701
776
735
841
771
830
596
732
1022
754
1336
1409
1568
1524
1480
1704
809
741
977
1240
1522
599
774
700
708
689
695
UHC
-------
TABLE A-2. DATA SUMMARY - COAL/OIL MIXTURE (CONCLUDED)
CO
CD
Test
No.
223g
h
i
j
k
1
m
n
0
P
q
r
224a
b
c
d
e
f
9
h
i
j
k
1
225a
b
c
d
226a
b
c
d
e
f
T25
2003
2038
2082
2067
1979
1934
1955
1963
1843
1879
1807
1923
1842
2134
2194
2138
2087
2196
2115
2129
2041
1852
2216
2189
2168
2223
2277
2222
T26
1611
1661
15,82
1631
1592
1509
1540
1489
1617
1587
1574
1553
1549
1472
1458
1552
1551
1458
1595
1507
1400
1274
1664
1669
1664
1783
1754
1754
1806
1844
1807
T27
792
819
812
778
805
772
777
759
799
789
712
719
682
626
629
681
671
629
669
676
670
595
895
924
960
995
1014
1040
998
987
1010
°2
3.9
3.3
3.9
3.7
3.2
3.9
3.6
3.8
4.1
3.4
3.5
3.3
3.6
2.8
2.9
3.6
3.2
4.1
3.6
3.1
3.8
7.6
4.0
3.7
4.5
3.7
5.1
6.2
3.5
3.8
4.9
CO
1156.6
122.7
87.1
625.3
161.8
426.1
123.1
116.8
229.0
190.9
68.0
84.0
131.0
687.0
134.1
96.7
169.8
87.0
119.7
95.2
109.0
136.7
138.9
136.5
194.8
64.0
74.0
121.0
C02
14.4
14.9
14.4
14.1
14.7
14.1
14.2
14.0
13.7
13.9
14.5
14.6
14.2
14.6
13.9
14.2
14.3
13.8
14.3
14.8
11.1
15.4
15.6
14.9
13.4
12.7
11.9
13.6
13.3
12.6
NO
219
234
256
221
221
611
336
712
524
733
547
581
568
246
168
381
480
810
205
427
238
1092
1159
1226
1156
1199
1266
1158
1152
1237
so2
716
726
728
724
717
729
689
700
694
712
683
691
764
911
794
792
781
832
726
827
783
883
2552
2536
2553
1695
1479
1446
1350
1403
1298
UHC
-------
TABLE A-3. DATA SUMMARY - RDF TESTING
Test
No.
2Z7a
b
c
d
e
f
g
h
i
228a
b
C
d
e
f
g
h
i
j
229a
b
c
d
e
f
g
230a
b
c
d
e
f
g
231a
b
232a
b
233a
234a
235a
b
236a
237a
238a
239a
240a
241a
b
c
d
242a
243a
244a
245a
b
Load
Btu/hr x TO6)
1.5
1.0
1.5
1.0
1.0
- 1.5
1 0
1.5
Excess
Air
5
10
30
X5
10
30
30
10
5
20
5
10
30
5
10
30
30
10
5
20
20
5
10
30
30
10
20
30
10
5
30
10
5
5
10
20
20
20
20
20
20
20
20
20
20
20
20
30
10
20
30
10
20
10
20
Primary
Fuel
4at. Gas
Coal
_
Refuse*
No.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
4
4
4
4
4
4
4
2
2
2
2
2
1
1
3
4
4
3
1
2
2
2
2
2
2
2
2
2
-
r
Refuse
Concen.
5
5
5
10
10
10
20
20
20 .
10
5
5
5
10
10
10
20
20
20
10
10
10
10
10
20
20
10
10
10
10
20
20
20
5
5
10
20
20
20
20
20
20
10
10
10
20
20
20
30
30
30
20
10
10
T
Preheat
Temp
290
300
310
310
315
315
310
315
300
300
300
300
300
300
300
300
300
303
300
290
310
310
316
313
317
310
304
305
313
313
314
319
319
470
560
300
300
302
302
302
309
300
317
319
319
309
304
300
300
300
300
300
300
317
315
Yaw
+6
Level
of Stack
Testing
\
3
2
2
2
2
2
2
3
3
3
3
3
-
-
2
3
3
3
3
3
1 = Ames
2 = Richmond
3 = Americology
4 = San Diego
181
-------
TABLE A-3. DATA SUMMARY - RDF TESTING (CONCLUDED)
Test
No.
227a
b
c
d
e
f
9
h
i
228a
b
c
| d
e
f
9
h
i
j
229a
b
c
d
e
f
9
230a
b
c
d
e
f
g
231a
b
232a
b
233a
234a
235a
b
236a
237a
238a
239a
240a
241a
b
c
d
242a
243a
244a
245a
b
T25
2066
2107
2149
2223
2172
2229
2200
2277
2334
1977
2231
2228
2251
2251
2301
2273
2155
2282
2320
1904
2055
2245
2281
2253
2257
2296
1987
2087
2093
2293
2269
2303
2326
2157
2304
2231
2311
2320
2054
1915
2035
2078
1976
1982
2139
2082
2017
2052
2066
2289
2196
1996
3452
T26
1802
1843
1885
1927
1892
1955
1958
2011
2049
1721
1956
1956
1969
1969
2011
1999
1920
2007
2047
1663
1795
1971
2006
2002
2022
2041
1750
1850
1858
1989
2000
2015
2034
1904
2014
1919
1966
1977
1751
1653
1759
1801
1757
1766
1931
1898
1831
1817
1865
1972
1720
1861
1995
T27
2092
2162
2217
2285
2262
2333
2298
2350
2396
2102
2344
2350
2370
2374
2411
2395
2302
2368
2426
2018
2195
2334
2381
2363
2360
2386
2081
2160
2175
2368
2351
2377
2395
2340
2460
2318
2429
2386
2039
1958
2050
2110
1995
1928
2117
2059
2068
2138
2165
2175
1964
2088
2328
Emission Level
CO
(ppm)
47.8
39
66
147.
31.2
50.8
88.9
972.3
29.4
53.7
715
42.4
1068
34.6
45.6
66.3
65.8
350
104
102
107
114.1
154.5
160.1
123.6
0.6
4.5
5.4
6.4
19.1
32.9
10.0
145.1
94.0
24
19
21
11
12
12
5
9
5
6
18
18
90.4
56.6
66
90
85
112
16
C02
(ppm)
9.4
11.3
9.6
12.5
11.9
9.7
9.9
13.0
10.5
11.0
11.4
8.8
12.2
10.8
9.3
9.8
11.2
12.1
10.2
10.3
11.4
10.8
9.2
9.7
11.7
9.5
8.7
9.9
10.9
8.5
10.5
10.7
17.7
17.8
10.8
11.5
12.9
11.5
9.7
11.0
11.3
11.4
10.6
11.3
15.0
14.9
15.9
15.8
13.8
16.0
15.3
16.6
14.4
NO
(ppm)
80
81
116
85
100
143
163
117
112
90
88
104
144
108
116
193
135
110
106
82
99
102
114
144
189
147
93
150
111
98
182
129
115
348
458
123
134
117
130
122
215
161
185
188
129
410
405
456
334
383
427
289
381
348
493
SO
(ppm)
59.7
168.7
169.6
167.2
165.5
183.2
3.1
9.8
9.8
14.8
22
24.9
30.7
8.4
34.7
54
UHC
(ppm)
1.0
5.0
8.0
4.0
5.0
4.0
6.0
6.0
6.0
7.4
6.1
13.7
13.1
15.1
23.0
34.1
10.5
10.7
19.0
19.9
34.9
29.1
23.5
2123.5
1902
6.6
18.6
22.1
22.1
34.8
35.9
51.4
18.1
1358.6
1321.1
1321.1
1243
1487
1575.3
1508.6
1148.6
1565.9
1692.8
4.0
3.8
4.5
4.0
4.0
3.8
0.5
11.2
0.4
1.3
436
°2,
('/)
3.8
2.0
4.4
1.2
2.2
5.6
5.2
i.o-
3.5
1.4
1.8
5.2
1.1
2.1
5.0
5.4
2.1
1.1
3.8
3.7
1.3
2.0
5.1
5.1
2.1
3.9
5.3
2.3
1.1
5.2
1.8
1.1
1.2
1.9
3.5
3.9
3.8
3.8
3.8
3.8
3.8
3.6
3.5
3.5
2.1
3.2
5.2
1.9
4.0
2.0
3.7
182
-------
COAL/OIL MIXTURE (COM) SUBSCALE COMBUSTION TEST RESULTS
CONDUCTED IN THE EPA/ACUREX MULTIFUEL FURNACE FACILITY
183
-------
INTRODUCTION
Subscale combustion tests with coal/oil mixtures as fuel were per-
formed by Acurex in the EPA Multifuel Furnace Facility to provide design
support for the planned full-scale COM facility at Lorillard Division,
Loew's Theaters, Inc., Danville, Virginia. The test objectives were as
follows:
Determination of emissions for 30 to 50 percent coal in No. 6
oil using identical fuels as anticipated for use at the Lorillard
demonstration site
Identification of fouling, piping, and pumping problems resulting
from fuel handling and combustion
t Determination of suitability of the Carbonoyl, Inc., COM additive
planned for use in the full-scale demonstration program
The subscale combustion tests consisted of two major activities:
Fuel preparation
t Combustion tests
These activities are described in the following sections.
1. FUELS AND FUEL PREPARATION
COM fuels for combustion testing were prepared with the coal and oil
identical to those anticipated for use at the Lorillard demonstration site.
No. 6 oil which meets Lorillard specifications and is identical to that
which is presently in use was obtained from Amerada Hess Corporation. The
high volatile bituminous coal which was determined to have the most desir-
able properties for wet grinding (from subscale wet grinding tests at Colo-
rado School of Mines Research Institute) and which will be used during demon-
stration testing was obtained from Maryland Coal and Coke Company. Specifi-
cations and chemical analyses of the oil and coal are presented in Tables 1
and 2.
Although a wet grinding ball mill will be used for demonstration
fuel preparation, dry grinding and subsequent mixing were used for the test
fuels. This preparation scheme was chosen because a suitable wet grinding
system was not available. Pulverized coal prepared by C$MRI was blended
184
-------
TABLE 1. NO. 6 OIL ANALYSES
Specifications3
API gravity
Sulfur (X Wt)
Flash point (PMCC °F)
Viscosity (SSF 0 122°F)
Pour point (F)
BS&W (X Vol )
15.3
2.22
204.0
247.0
+50.0
0.4
Ultimate (
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
% Wt)a
84.71
10.75
0.36
1.93
2.22
0.03
Supplied by Amerada Hess Corporation
TABLE 2. COAL ANALYSES
Proximate (% Wt)a
Ultimate (% Wt)b
Moisture 4.2
Volatiles 33.0
Fixed carbon 54.0
Ash 8.9
Ash fusion temp (F) 2700.0
Hardgrove grindability 68.0
Btu per pound 13368.0 As Rec'd
13954.0 Dry
Origin: Clintwood seam, Conoway, Virginia
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen
79.0
5.0
1.5
0.9
13.4
Supplied by Maryland Coal and Coke Company
'EPA-650/2-75-046, May 1975
185
-------
with No. 6 oil and the Carbonoyl additive in a high turbulence batch mixer
supplied by Littleford Brothers. The grind distribution of the coal was
approximately 80 percent passing 200 mesh and 100 percent passing 48 mesh.
The additive was prepared in a 50-percent aqueous solution and constituted
3.75 percent by weight of the COM independent of coal fraction.
The blending procedure was as follows:
1. Place premeasured No. 6 oil in the Littleford Brothers batch
mixer (Model FM 13100 20-gallon capacity). The mixer is main-
tained in the "on" position. Mixer is steam jacketed and mix-
ture is maintained at about 140°F.
2. Add premeasured additive solution to oil
3. Add premeasured pulverized coal to oil and additive and allow to
mix for 10 minutes
4. Discharge into 55-gallon storage drum
Fuel mixing occurred between July 11 and July 22, 1977. Approxi-
mately 1500 total gallons of 50-, 40-, and 30-percent COM were prepared.
No unexpected difficulties arose. Those problems which did occur were
related to handling of the fuels, particularly the pulverized coal. None
of the handling problems, however, are related to full-scale operation.
The COM was stored at ambient temperatures (minimum approximately
50°F) for up to 24 days before use. About 3 hours prior to use, the
storage drum was wrapped with electrical resistance heating blankets and a
mixer with a 6-inch propeller was immersed in the mixture. During this
period, the mixture temperature rose to about 140 to 150°F. A homogeneous
mixture was observed at about 100°F. The mixture was pumped into tanks
located within the facility. The empty storage drums were examined for
signs of pulverized coal which had settled in the mixture during storage
and failed to reentrain during the mixing cycle. In all cases, no deposits
of pulverized coal were found.
2. SUBSCALE COMBUSTION TESTS
Subscale combustion testing occurred between July 27 and August 11
at the EPA/Acurex facility. The test facility, shown in Figure 1, is
186
-------
OBSERVATION
PORT
00
PARTICULATE SAMPLING PROBE
GASEOUS EMISSION SAMPLING PROBE
FURNACE WATERWALLS
REFRACTORY
SECONDARY AIR
LINES (5)
CONVECTIVE
TUBE BANK
BURNER
3 MILLION
BTU/HR
AXIAL
FUEL TUBE
POSITION
CONTROL
\ /
VARIABLE
SWIRL
CONTROL
V
CONVECTIVE TUBE
OBSERVATION PORT
FLAME
OBSERVATION
WINDOW
ASH PIT
Figure A-1. EPA/Acurex Multifuel Furnace Test Facility - 3 million Btu/hr
capability side view. (See Figure 2 for Section A-A)
-------
sponsored by the Environmental Protection Agency to investigate advanced
emission control concepts for utility and industrial boilers. The configu-
ration additions indicated on the sketch were made to more closely simulate
industrial boiler operating conditions. One important addition to these
tests was the steam-cooled tube bank across the path of the combustor gases.
The purpose of these tubes was to model the convective section of a boiler
and thereby provide information regarding tube fouling. Water-cooled tubes
spiralled around the combustion chamber were added to simulate the water-
walled combustion chamber of industrial watertube boilers.
2.1 FACILITY MODIFICATIONS
2.1.1 Convective Tubes (Slagging Probes)
A.bank of four tubes mounted across the gas flow was designed to
simulate the entrance plane of the convective tube banks of the demonstra-
tion boiler with the primary objective of gaining qualitative information
regarding the fouling tendencies of the Lori Hard fuels. The tube con-
figuration is shown in Figure 2. The tube sizing and spacing were selected
to duplicate the velocity through the tubes of the demonstration unit.
Cooling was provided to maintain the tubes below 600°F, the factory esti-
mated temperature of the convective tubes.
2.1.2 Combustion Chamber Waterwalls
As shown in Figure 1, the combustion chamber preceeding the convec-
tive tube section was lined with several loops of copper tubing for radiant
cooling. This cooling reduced the bulk gas temperature to below 2300°F
which is the factory estimate of gas temperature entering the convective
section of the demonstration boiler. The cooling loops were in the three
horizontal extension sections.
2.1.3 Fuel Supply System
The fuel supply system is shown schematically in Figure 3. The item
numbers shown are described in an equipment list in Table 3.
188
-------
INNER FURNACE
DIAMETER 30"
V SQUARE
REFRACTORY
SEE INSERT BELOW
FOUR EQUALLY
SPACED, STAINLESS
STEEL TUBES 1-1/2"
DIAMETER
TUBE SPACING
13/16"
TWO THERMOCOUPLES PER
TUBE (ON FLAME SIDE)
INSERT
COOLING AIR AND WATER OUT
COOLING AIR IN
WATER INJECTION
Figure A-2. Convection tube bank.
(Section A-A from Figure A-l)
189
-------
MXJ-tXJ-KXMxH
N-14
NOZZLE
Figure A-3. Fuel supply system schematic.
-------
TABLE 3. FUEL SUPPLY SYSTEM EQUIPMENT LIST
Item
T-l
T-7
T-9
M-3
M-10
M-ll
HT-2
HT-8
HT-9
TE-101
TE-102
TE-103
TE-106
P-4
P-5
TIC-102
TIC-103
TIC-106
F6A
F6B
TI-104
TI-107
TI-108
H-12
PI-107
PI-108
V-l
V-2
V-3
V-4
N-14
FCV-105
Description
Fuel storage tank
Fuel holding tank
Fuel holding tank
Pneumatic mixer, propeller type
Pneumatic mixer, propeller type
Pneumatic mixer, propeller type
Heating blanket
Strip heaters
Strip heaters
Temperature element
Temperature element
Temperature element
Temperature element
Gear pump
Helical rotor pump
Temp indicating controller
Temp indicating controller
Temp indicating controller
Strainers
Strainers
Temperature indicator
Temperature indicator
Temperature indicator
Circulation heater
Pressure indicator
Pressure indicator
Metering valve
Solenoid valve
Ball valve
Ball valve
Nozzle
Flexible control valve
Comments
55-gal drum
135-gal
135-gal
1.3 hp, variable speed
1.0 hp, variable speed
1.0 hp, variable speed
1200 W
8 per tank, 500 W each
8 per tank, 500 W each
1.0 hp, 450 RPM
0.75 hp, 1200 RPM
70 to 250°F
70 to 250°F
70 to 250°F
1/16" perforations
1/16" perforations
60 to 260°F
60 to 260° F
60 to 260° F
3000 W
0 to 160 psi
0 to 160 psi
Self cleaning
Flame safety
Sampling port
Nozzle flow recirc
Regulates recirculation
191
-------
For transfer of fuel into the holding tanks, (Item Tl, Table 3} a
COM storage drum is preheated using electrical resistance heaters (Item HT-2)
and agitated with a pneumatically driven shaft mixer (Item M-3). The drum
is connected by flexible hose to the system inlet where the mixture is
pumped through either the gas pump (Item P-4) or the helical rotor pump
(Item P-5) (or both pumps) to holding tank 1 or 2 (Items T-7 and T-9).
During furnace operation, COM is pumped from either holding tank
through one of two parallel strainers (Items F6A or F6B) by either or both
pumps. The flow then splits into a recirculation line and a nozzle line.
The recirculation line returns excess flow to the holding tank. A flexible
pinch valve (Item FCV-105) regulates the recirculation flow and as a
result acts as a coarse adjustment for flow into the nozzle. COM flow to
the nozzle is directed through a circulation heater (Item H-12). The fine
adjustment on flowrate is done with a self-cleaning metering valve (Item
V-l). Temperatures and pressures are monitored on either side of this
valve (Items TI-107 and TI-108, PI-107 and PI-108). A solenoid valve
(Item V-2) is wired to the furnace flame safety system. Flow progresses
from the system outlet through flexible hose to the nozzle (Item N-14).
Fuel samples may be drawn at any time (Item V-3). Prior to light-off,
nozzle flow may be redirected into the recirculation line (Item V-4).
Several other flow options are available. COM may be transferred
from either of the holding tanks to the other holding tank or to the
original storage drum. Fuel may be delivered to the nozzle directly from
the storage drum. This operation requires that both pumps be employed,
however. The circulation heater (Item 38) may be used to augment holding
tank heaters prior to startup.
The entire system was electrically heat traced and insulated to
maintain COM temperatures to at least 140°F. Temperature was controlled by
five individual thermostats covering the storage tank to delivery system
inlet line; the holding tanks, system piping and the circulation heater.
These on-off type thermostats were capable of controlling fluid temperature
from 60°F to 250°F with a 7°F tolerance.
A duplex pumping arrangement was chosen such that system shutdown
would be prevented in case one pump failed and also to compare operation
192
-------
of the two pumps on COM. The fuel flow through each pump was approximately
8 gpm yielding fluid velocities in the 3/4-inch lines of approximately
5 ft/second.
2.1.4 Atomization Air System
Standard shop air at 150 psig at flowrates up to 30 scfm was used
for fuel atomization. The pressure and flowrate at the nozzle were con-
trolled by appropriate pressure regulators and flowmeters.
2.2 EMISSION MONITORING EQUIPMENT
Continuous monitors were used to collect emission data. Table 4
lists the instrumentation used and the principle of operation of each
device.
Calibration was performed prior to, during, and at the end of each
test period. Correction in the emission data due to calibration shifts,
whenever present, were taken into account in the calculation of reported
NO and CO levels.
2.3 CHECKOUT TESTING
Prior to actual testing, several hours of system checkout and equip-
ment evaluation were conducted.
During operation on natural gas, it was discovered that the convec-
tive tube bank was not adequately cooled by air alone, and a water injec-
tion system was added (see Figure 2). The goal was to maintain the tubes
at about 600°F, but the coarse control afforded by the water injection
provided temperatures of about 400°F. Also, the water in the tubes
caused differential expansion between the top and bottom of each tube.
As a result, tube-to-tube and tube-to-wall spacing changed during the
test run.
Two nozzle configurations were evaluated on No. 6 oil and 30-
percent COM during checkout. A Delavan Corporation swirl air nozzle and
a Sonic Development Corporation Sonicore nozzle were tested. The Delavan
nozzle, designed for 60 gph maximum flow with a 70-degree spray angle
performed well with moderate burner secondary air swirl. No flame
193
-------
TABLE 4. EMISSION MONITORING INSTRUMENTATION
Pollutant
NO
so2
CO
co2
°2
Parti cul ate
Loading
Type of Operation
Chemi 1 umi nescence
Pulsed Fluorescent
Nondispersive
Infrared (NDIR)
Nondispersive
Infrared (NDIR)
Paramagnetic
Cyclone and
Filtration
Manufacturer
Ethyl Intertech
Thermoelectron
Ethyl Intertech
Ethyl Intertech
Ethyl Intertech
Acurex Corp
Models
Air Monitor-
ing
Teco
Model 40
Uras 2T
Uras 2T
Magnos 5A
HVSS
Instrument
Range
0-5 ppm
0-10
0-100
0-250
0-1000
0-5000
0-50 ppm
0-100
0-500
0-1000
0-5000
0-500 ppm
0-2000
0-52
0-20%
0-5%
0-21%
0-3 ym
Minimum
-------
impingement on the furnace walls was observed. Fuel pressure of about 40
psi delivered the desired 10 to 15 gph. An atomization air pressure of 40
psi and a flowrate of 1000 scfh adequately atomized the 180 to 200°F fuel.
When the fuel temperature was greater than about 200°F, pulsations in the
flame were observed. These pulsations, thought to be due to the vaporiza-
tion of water in the fuel, stopped when the fuel temperature was lowered.
Since atomization was adequate at 180 to 190°F, the remainder of the tests
were run in this temperature range. Following approximately three hours of
operation on 30-percent COM, extensive erosion was observed on the nozzle
tip. Figure 4 shows the nozzle and the areas where erosion occurred. The
erosion was probably a result of the high fuel-air mixture velocities (600 to
1000 feet per second) necessary for proper atomization. Although the
part was supplied as stainless steel, it was discovered later to be
carbon steel. A stainless steel metering nut with a tungsten carbide pintle
was then obtained to minimize erosion.
Figure 5 shows a diagram of the Sonicore nozzle. Designed for flow-
rates up to 60 gph, it was operated at fuel pressures of 0 to 5 psig at the
nozzle and atomization air pressure and flowrate of 35 psig and 1500 scfh,
respectively. Initial nozzle operation at moderate and low secondary air
swirl resulted in clinker formation at the nozzle tip within 15 minutes of
light-off. This condition was eliminated at a zero swirl setting, but the
flame was unstable. Following approximately 3 hours of operation on 30-
percent COM at various swirl levels, examination of the nozzle did not
reveal any erosion. A decision was made to use the Delavan nozzle. Even
though the Sonicore nozzle was more erosion resistant, the Delavan nozzle
provided superior flame characteristics.
During the checkout tests, a thermocouple mounted adjacent to the
furnace wall just prior to the convective tube bank was used to estimate
combustion gas temperatures when the suction pyrometer failed. Also, a pro-
portional controller was used to minimize fluctuations in fuel flowrate
caused by the "on-off" characteristic of the circulation heater thermostat.
The flow variation resulted from the temperature-induced change in the
viscosity.
195
-------
MIXING CHAMBER
AIR INLET
FUEL INLET
AIR INLET TO
MIXING CHAMBER
EROSION
OCCURRED
IN THESE
AREAS
PINTLE PLATE
METERING NUT
Fiyure A-4. Dcliiviin swirl-air nozzle
196
-------
RESONATOR CHAMBER
FUEL
STANDING SHOCK WAVE
AIR
Figure A-5. Sonic Corporation Sonicore nozzle.
197
-------
2.4 TESTING
The test points completed are shown in the matrix of Table 5
TABLE 5. COM TEST MATRIX
2 Excess
Air
% Coal
20
30
40
1.8 Million Btu/hr
30
40
50
100
1.35 Million Btu/hr
30
40
50
100
For a furnace load of 1.8 million Btu/hr corresponding to the heat
release rate of the demonstration boiler at full load (80,000 pph), three
coal-oil mixtures were burned for a range of excess air. To provide
reference points, 100-percent No. 6 oil and 100-percent pulverized coal
were fired at the same conditions. Additional data was taken at a reduced
load of 1.35 million Btu/hr corresponding to 60,000 pph for full-scale.
Test conditions at 40 percent of full-scale load (0.72 million Btu/hr) were
planned but eliminated because of the limited range of burner secondary air
control.
2.4.1 Test Narrative
Tests were first performed with 100 percent pulverized coal at a
firing rate of 1.8 million Btu/hr. Following nearly 3 hours of operation,
approximately 50 percent of the convective passages were blocked by ash
deposits. The hard, porous deposit was removed mechanically.
A case hardened (Rc = 58 to 0.03 inches) Delavan nozzle was used to
obtain test points on No. 6 oil. This was a higher capacity nozzle (100 gph)
than had been used in the checkout runs, but no differences were observed
either in flame shape or in emission levels. After 3 to 5 hours of testing,
inspection of the nozzle showed no signs of erosion. Also, no ash deposi-
tion on the convective tube bank was noted.
198
-------
A 30-percent COM was tested next using the case-hardened nozzle used
on the No. 6 oil. After about 3 hours of operation, significant deteriora-
tion of the flame shape suggested nozzle tip erosion which was confirmed by
inspection. The erosion pattern was similar to the first eroded nozzle
(see Figure 4); substantial erosion occurred on the pintle plate with
slight but definite erosion occurring on the metering nut. The nozzle life
was unaffected by the case hardening process. A new nozzle with tungsten
carbide pintle and stainless steel metering nut was used to complete the
tests. This 60 gph nozzle was compared with the larger nozzle by taking two
duplicate points; no difference was observed in the two nozzles. After
approximately 2-1/2 to 3 hours, the nozzle was inspected and significant
erosion was observed. In this case, only the stainless steel metering nut
had eroded and the tungsten carbide pintle remained unchanged.
The convective tubes showed some fouling but this was minimal
compared to the pulverized coal.
A 50-percent COM was tested next. A high capacity (150 gph) Delavan
nozzle was chosen to minimize velocities through the nozzle tip. Test data
was taken prior to flame deterioration at the 3-hour point when erosion was
again observed. At this point, approximately 10 percent of the total con-
vective passage was occluded by ash deposition. Factors contributing to this
high rate of deposition were probably the volumetric heat release (0.75
x 106 Btu/hr-ft3) and the hot refractory wall of the furnace.
The 40-percent COM was the last fuel tested. Equipment problems
arose after about 1 hour of testing. First, the helical rotor pump failed.
Subsequent inspection revealed that the rotor was extensively galled. The
gear pump was used for the remainder of the tests. Second, plugging of the
fuel metering valve was experienced. The valve was removed and cleaned
twice without success.
After flushing the entire system with No. 6 oil, a second attempt at
40-percent COM was made. Similar valve plugging was experienced. The valve
was replaced with a conventional needle valve and data points were taken.
Erosion of the nozzle was observed following these tests, but 1t was less
than that leading to eratic flame patterns. Fouling was less than with
199
-------
50-percent COM but greater than with 30-percent COM. Again, the high
volumetric heat release and the hot-wall effect probably contributed to the
high rate of deposition.
2.4.2 Emission Tests
Pollutant emissions for each fuel were measured at excess air levels
of 20-, 30-, and 40-percent. For these measurements, the furnace load was
maintained at 1.8 x 106 Btu/hr. This load corresponds to a volumetric heat
release of 0.75 x 106 Btu/hr-ft3, which is approximately the same as the
demonstration boiler at full load.
2.4.2.1 Nitric Oxide (NO) Emissions
Figures 6 through 10 show NO levels as a function of stoichiometric
ratio for No. 6 oil, 30-, 40-, and 50-percent COM, and pulverized coal. In
Figure 8, the data taken in test 201a is believed to be most representative
as it was taken prior to plugging problems experienced with the fuel supply
system. The data recorded in tests 201b, c, and d, are questionable since
partial fuel supply blockage occurred during these tests. As expected, NO
emissions increased with stoichiometric ratio for each fuel. This was
probably due to increased oxidation of fuel nitrogen with each increase in
excess air. The rate of increase of NO with excess air was greater for coal
and coal-oil mixtures than for oil alone. This is attributed to the in-
creased emissions of fuel NO for the coal-containing fuels. Fuel NO is
generally more sensitive to excess air levels than thermal NO which pre-
dominates with oil combustion. Table 6 lists general properties of the
five fuels tested. Note that fuel nitrogen increases as coal content of
the fuel (mixture) increases.
The effect of coal content on NO emissions is shown in Figure 11.
The upper curve represents NO levels recorded at 40 percent excess air,
and the lower curve represents data taken at 20 percent excess air. NO
emissions from COM combustion were slightly lower than levels expected
from a straight proportional weighting of emissions according to weight
percentage of coal and oil.
200
-------
ro
a.
O
BOO
700
eoo
500
400
3OO
200
100
A
FUEL: NO. 6 OIL
BURNER: 3 MILLION BTU/HR IFRF; OELAVAN NOZZLE
FURNACE CONFIGURATION: HORIZONTAL EXTENSION
SWIRL: 6 OF 8
LOAD: 1.8 MILLION BTU/HR
TEST ID SYMBOL % EXCESS AIR
198a
198b
198c
198d
1986
198f
1.0 1.1 1.2 13 1.4 1.5 1.6 1-7
STOICHIOMETRIC RATIO
20
20
30
40
10
40
Figure A-6. Nitric oxide (NO) versus stoichiometric ratio: No. 6 oil,
-------
ro
o
Q.
Q.
O
MO
700
eoo
500
400
300
200
100
FUEL: 30 PERCENT COM
BURNER: 3 MILLION BTU/HH IFRF; DELAVAN NOZZLE
FURNACE CONFIGURATION: HORIZONTAL EXTENSION
SWIRL: 5 OF 8
LOAD: 1.B MILLION BTU/HR
TEST ID SYMBOL % EXCESS AIR
199a
199b
199c
O
Q
O
A
199e
199f
20
30
30
20
30
40
"1.0 1.1 12 13 1.4 1.5
STOICHIOMETRIC RATIO
1.6
1.7
Figure A-7. Nitric oxide (NO versus stoichiometric ratio: 30-percent COM.
-------
ro
o
oo
I
Q.
800
700
800
500
400
300
200
100
FUEL: 40 PERCENT COM
BURNER: 3 MILLION BTU/HR IFRF: DELAVAN NOZZLE
FURNACE CONFIGURATION: HORIZONTAL EXTENSION
SWIRL: 5 OF 8
LOAD: 1.8 MILLION BTU/HR
TEST ID SYMBOL
201a
201 b
201C
201 d
o
GJ
2
EXCESS AIR
20
20
30
20
FLAGGED SYMBOLS REPRESENT
QUESTIONABLE DATA DUE TO
FUEL SUPPLY SYSTEM PLUGGING.
1.0 1.1 1.2 1.3 1-4 1.5 1.6
STOICHIOMETRIC RATIO
1.7
Figure A-8. Nitric oxide (NO) versus stoichiometric ratio: 40-percent COM.
-------
0.
EXCESS AIR
20
30
40
1.0 1.1 1.2 1.3 1.4
STOICHIOMETRIC RATIO
1.5
1.6
1.7
Figure A-9. Nitric oxide (NO) versus stoichiometric ratio: 50-percent COM,
-------
1200
1000
ro
o
en
O
800
800
400
200
1.0
11 1.2 1.3
STOICHIOMETRIC RATIO
1.4
FUEL: PULVERIZED COAL
BURNER: 3 MILLION BTU/HR IFRF; B«W SPREADER
FURNACE CONFIGURATION: HORIZONTAL EXTENSION
SWIRL: 4 OF 8
LOAD: 1.8 MILLION BTU/HR
TEST ID
197«
1976
197c
1970
197e
O
Q
% EXCESS AIR
20
30
40
10
5
1.5
PREHEAT TEMP.
100
100
102
102
102
Figure A-10. Nitric oxide (NO) versus stoichiometric ratio: pulverized coal
-------
TABLE 6. FUEL PROPERTIES
Fuel
Type
No. 6 oil
30% COM
40% COM
50% COM
100% Coal
% N
By Weight
0.36
0.70
0.82
0.93
1.5
% S
By Weight
2.22
1.82
1.69
1.56
0.9
% Water
By Weight3
0.0
4.82
5.24
5.66
4.2
% 02
By Weight
0.0
2.67
3.56
4.45
8.9
HHV
Btu/lb
18,800
17,170
16,627
16,084
13,368
Includes 3.56 percent of fuel by weight due to fuel additive.
206
-------
ro
1400
1200
1000
800
600
400
200
O 20% EXCESS AIR
A 30% EXCESS AIR
G 40% EXCESS AIR
40% EXCESS AIR
20% EXCESS AIR
OIL
ONLY
10 20
30 40 50 60
PERCENT COAL IN FUEL
PULVERIZED
COAL ONLY
Figure A-ll. Nitric oxide (NO) versus percent coal in fuel
-------
The lower NO emissions of the COM fuels could have been caused by
the water content of the additive, or the shielding of coal by the oil
spray. Water content in the fuel has been found to reduce NO emissions.
During some studies of water emulsions with distillate and residual oils
the NO levels were reduced an average of 100 ppm when approximately 5-
percent water was added to the fuel oil.* In the COM tests the water con-
tent of the fuels attributed to the additive were 3.56 percent by weight.
Accounting for this water content, on the basis of the emulsion tests cited
above, yields NO levels which conform more closely to proportional levels
based on No. 6 oil and pulverized coal NO emissions. The second factor
which may have affected NO emissions during these tests results from the
layer of oil surrounding each coal particle. This oil layer delays oxygen
diffusion to the coal and suppresses oxidation of coal fuel nitrogen to NO.
2.4.2.2 Carbon Monoxide (CO) and Unburned Hydrocarbon (UHC) Emissions
Carbon monoxide (CO) and unburned hydrocarbon emissions were nearly
zero for all COM, oil and coal tests. CO emissions were insignificant even
though the excess air levels were reduced to 10 percent on occasion. UHC
emissions were undetectable during all tests.
2.4.2.3 Particulate Mass Loading
Particulate stack sampling tests were conducted for the 50-percent
coal and pulverized coal fuels. The results indicate that very low frac-
tions of the ash contained in the fuel went out the stack. Based on the
relatively small amounts of ash remaining in the furnace (compared to the
amount of fuel fired), the stack test results are in question. The test
results are also contradicted by more extensive testing at General Motors
where nearly 100 percent of the ash appeared in the flue gas.
*G.B. Martin, "Evaluation of NOX Emission Characteristics of Alcohol Fuels
in Stationary Combustion Systems," presented at the Joint Meeting Western
and Central States Sections, The Combustion Institute, San Antonio, Texas,
April 21 to 22, 1975.
^Brown, A, "First Report of the General Motors Corporation Powdered Coal-In-
Oil Mixtures Program," ERDA Contract E(49-18)-2267, December 1976.
208
-------
2.4.3 Discussion
Pumping of the COM resulted in the failure of the helical rotor pump
after approxiamtely 50 hours of operation. On disassembly, no damage was
observed to the BUNA-N stator. However, significant galling was seen on the
lobes of the rotor. Two possible explanations for this failure are: the
pump should have operated at 400 to 500 rpm but instead operated close-
coupled to a 1200 rpm motor; the pump operated in a dry state for short
periods. The transfer of COM from the storage tank to the system feed tank
was probably responsible for the dry operation of the pump. For the transfer
operation, the helical rotor pump was actuated as soon as the mixture was
thought pumpable. This resulted in dry pump operation if the mixture was
not flowing.
The Viking gear pump did not fail, but developed leaks in the packing.
When disassembled after test completion, hardened or congealed fuel was
detected between the shaft and the packing resulting in leakage. This prob-
lem could probably be remedied by the use of mechanical seals. Inspection
of the pump components susceptible to wear showed no indication of abrasion
after about 50 hours of operation.
Metering problems resulted primarily from the very low flowrates
associated with test conditions. The self-cleaning micrometering valve had
a maximum fuel passage dimension of about 0.125 inches. Operation was found
to be most effective if the valve metering position was between 75 and 100
percent open. Operation in this range was possible by controlling the
amount of recirculation by adjusting the flexible valve. While the valve
performed satisfactorily with both 30- to 50-percent COM, excessive plugging
occurred with the 40-percent COM. For the 40-percent COM, however, operation
was impossible even in the 100-percent open position. The plugging occurred
upstream of the metering groove where the self-cleaning feature was ineffec-
tive and where plugging would not be suspect because of its larger dimensions
(approximately 0.25-inch).
No explanation can be given for the failure of the metering valve on
the 40-percent COM while not on the 50-percent COM. Subsequent analysis of
209
-------
fuel samples indicated a +2 percent tolerance on the total solids content
of the fuels indicating that the difficulty was not in improperly prepared
fuel.
Although the exact cause of the failure is not known, two definite
possibilities exist.
Insufficient additive in the fuel resulted in particle agglomera-
tion leading to eventual plugging
Following each COM fuel, the entire system was flushed with No. 6
oil stored in the second holding tank. Continued flushings
resulted in a mixture of No. 6 and the coal in the second holding
tank and in the system lines. In standing for any period of
time, this coal would settle onto the walls of the pipes. Settle-
ment occurred because the dilution of the additive rendered it
ineffective. Agglomerated coal particles would then reentrain
from the pipe walls when flow occurred.
Of these two failure modes, the second is the most suspect since each
drum of fuel required two and a half batches of fuel preparation. Carbonoyl
Company has stated that little difference in fuel characteristics would be
observed if the additive was not included in one of these batches. Also, it
is highly improbable that three consecutive batches would have been made
without additive. The second failure mode is supported by the fact that
plugging was the result of agglomerated coal particles. The packing of
these particles was very similar to that observed when unstabilized coal is
removed from a sample after settling.
Significant erosion of the Delavan nozzle was experienced. At the
time, this occurrence was viewed as a major problem but further investiga-
tion indicated that a change in nozzle design would probably remedy the
problem. The Delavan nozzle uses an impingement type internal-mix atomiza-
tion scheme which, by design, is highly susceptible to erosion. A recommen-
dation for future use of air atomizing nozzles are the external-mix type
where atomization takes place after the fuel and air have left the nozzle.
Several types under this general design are available.
210
-------
Fouling of the convective tube bank was observed for all coal-con-
taining fuels. Deposition was greatest, as would be expected, for the
pulverized coal and least for the 30-percent COM. For the pulverized coal,
about half of the convective flow passages were occluded. Deposition
associated with the 50-percent COM was less obvious. Due to the displace-
ment of the convective tubes, some tube-to-tube and tube-to-wall spacings
were reduced. Increased deposition was noted for these spacings while
generally no fouling occurred in spacings which remained equal to or wider
than planned.
Two existing conditions which probably contributed substantially to
tube fouling were the volumetric heat release rate of the test points and
the loss of radiant cooling in the combustion chamber. The volumetric heat
rate of the tests corresponded to that of the demonstration boiler at full
load. Since the initial tests were made at this load, fouling had already
occurred when the load was reduced to 1.35 x 106 Btu/hr and any subsequent
slagging went unobserved. The loss of coolant tubes during checkout prob-
ably contributed to slagging as well. The radiating refractory was
sufficiently hot to melt the impinging ash particles. Slag thickness
averaged about 0.25 inch at the completion of the test program. These
slagging results indicate that tube fouling is a possibility at full-scale
operation, but firm conclusions are inappropriate at this time due to the
variation in results at both GM and PERC (Pittsburgh Energy Research Center)
211
-------
A A XNI lOCV additional routing
. Corporation
INTEROFFICE
to: Craig Derbidge
from: Allen Shimizu
date: September 15, 1977
The viscosity tests on the Lorillard fuels have been completed,
although their validity is in question. Doubt arises from the exceedingly
high viscosities of the 50% slurries. This indicates either improper
measurement procedure or extremely high viscosities. A study of the litera-
ture for operation of the Brookfield viscometer did not reveal any error
in procedure. In addition, prior to the viscosity measurements, the instru-
ment was calibrated with two standard liquids. This all points to a highly
viscous mixture at 50%.
Figure 1 shows viscosities for the following fuels:
GM oil: 472 coal weight COM
Lori Hard oil: 50% COM
t Chevron oil & Ptsbg. 18 oil: 30, 35, 40, 50% COM
t PERC oil: 20, 40% COM
Note that the PERC and GM oils are substantially less viscous than the Lori Hard
and Chevron oils. Also note that the slurries made with the GM and PERC oils
are far less viscous than those made with the Chevron oil. This leads one to
believe that the viscosity of the mixture may be a strong function of the oil
viscosity. Figure 2 plots the ratio of the mixture viscosity to the oil vis-
cosity versus the X coal 1n the mixture. This figure Indicates that the
"normalized viscosities" as a function of coal fraction at 150°F are similar
for the PERC 20 and 40% COM. the GM 47% and the Chevron 30. 35. and 40% COM.
The two points at 50% seem to be extremely hijjh, but this may be due to the
coal fraction. According to Brown at GM, the viscosities of the mixtures In-
creased significantly as the coal fraction approached and exceeded 50%. His
comments are reflected in the Chevron/Ptsbg. #8 tests showing the 30, 35
and 40% viscosities with uniformly Increasing values, while the 50% COM
exhibits anomolously high values.
Also shown on Figure 1 are viscosities of the GM oil and Marathon
oil mixtures made with the petrol ite additive. The GM oil and -200 mesh
coal 1s denoted by PGM 30 and PGM 50 for 30 and 50% COM. The Marathon oil.
which 1s similar In viscosity to the Chevron oil, 1s denoted MAR 30 and
"MAR 50. These were also made with -200 mesh coal. Those marked with FN
Indicate -325 mesh coal. This shows a slight viscosity Increase with finer
particles.
212
-------
September 15, 1977
Allen Shinrizu memo
COM Viscosity (con't)
Conclusions:
Our viscosity measurements were correct
§ Mixture viscosity 1s a strong function of oil viscosity
Viscosity of the mixture Increases significantly as the
coal fraction approaches and exceeds 50%
Although decreasing particle size Increases viscosity,
- viscosity appears to be only a weak function of particle
size
t GM used a #6 oil that was particularly fluid
The GM 46.6% COM should be questioned
ABS:mmcL
213
-------
10*
50% LORILLARD FUELS
10s
MAR 50FN
104
2
UJ
t-
1U
0
(0
o
103
PERCENT SOLIDS
PERC OIL
10'
80 100 120 140 160 ISO 200
TEMPERATURE - *F
Figure 1. Viscosity of various COM fuels.
214
-------
iooo r
LORILLARD
CHEVRON/W
TEMPERATURE = 150°F
GM + PETROLITE @160*F
MAR/CHEVRON @160*F
O 100
(D
6
O
O
10
O
/
*
V
10
20 30
PERCENT COAL
40
50
Figure 2. Relative COM viscosity versus coal
mixture ratio (percent by weight).
215
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-80-043
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Pilot Scale Combustion Evaluation of Waste and
Alternate Fuels: Phase III Final Report
5. REPORT DATE
March 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
R.A. Brown and C. F. Busch
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Acurex Corporation
Energy and Environmental Division
485 Clyde Avenue
Mountain View, California 94042
10. PROGRAM ELEMENT NO.
EHE624A
11. CONTRACT/GRANT NO.
68-02-1885
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Phase III Final; 2-8/78
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES IERL-RTP project officer is David G. Lachapelle, Mail Drop 65,
919/541-2236. EPA-600/7-79-132 was the Phase II final report; there was no Phase I
final report.
is. ABSTRACTThe report gives results of three studies at EPA's Multifuel Test Facility.
The first evaluated a distributed-air staging concept for NOx control in pulverized-
coal-fired systems. The results showed that minimum NO levels of 140 ppm were
achieved at overall residence times similar to those used during conventional sta-
ging tests. However, the NO levels achieved with the distributed-air concept were
no lower than those achievable with conventional staging. The second evaluated com-
bustion control techniques and NO emissions when firing coal/oil mixtures. NO emis-
sions for a given burner and nozzle were generally proportional to the fuel-nitrogen
content of the fuel. Additionally, combustion control technology currently used for
NOx control from pulverized coal was found to be effective with coal/oil mixtures,
but to differing degrees, depending on the coal/oil mixture ratios and compositions.
The third evaluated emissions and combustion characteristics of refuse-derived fuel
(RDF) co-fired with either natural gas or pulverized coal. Four RDF materials were
evaluated for gaseous, particulate, trace metal, and organic emissions. In general:
CO and UHC emissions were low; NOx and SOx emissions decreased with increasing
RDF content when co-fired with coal; particulate levels did not substantially increase
with the RDF; and no trace metal emissions correlation was found.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATl Field/Group
Pollution
Nitrogen Oxides
Combustion Control
Refuse
Wastes
Coal
Fuel Oil
Pollution Control
Stationary Sources
Staged Combustion
Refuse-derived Fuel
Coal/Oil Mixtures
Alternate Fuels
13B
07B
21B
21D
13. DISTRIBUTION STATEMEN1
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
227
2O. SECURITY CLASS (This page/
Unclassified
22. PRICE
IPA Form 2220-1 (9-73)
216
------- |