EPA- 600 / 7- 89- 0l5a
December 1989
' FIELD EVALUATION OF LOW-EMISSION COAL BURNER
TECHNOLOGY ON UTILITY BOILERS
VOLUME I
Distributed Mixing Burner Evaluation
A. R. Abele, G, S. Kindt, R. Payne
(Energy and Environmental Research Corporation)
and
P. W. Waanders
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44203
EPA Contract 68-02-3130
EPA Project Officer: P. Jeff Chappell
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
Prepared for:
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460

-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.
ii

-------
ABSTRACT
This report describes the results of a study in which N0X emissions and
general combustion performance characteristics of four burners were evaluated
under experimental furnace conditions. Of primary interest was the
performance of a low-NOx Distributed Mixing Burner (DMB), which was tested in
a nominal full-scale (120 x 10® Btu/hr or 35 MW) version and in a
corresponding half-scale version. Performance was compared against a half-
scale commercial 1 ow-KIOx Dual Register Burner (DRB) and a 120 x 10® Btu/hr
(35 MW) commercial Circular Burner. The report documents the performance of
each burner type over a wide range of firing conditions and for different
bituminous and subbituminous coal types.
Additional goals of the test program were to provide information
relating to the effects of burner design, burner scale and thermal
environment on N0X emission performance. Full- and half-scale DMB
performance was compared under equivalent thermal conditions; the DMB was
tested under two levels of furnace insulation; results from the DRB and
Circular Burners were compared to field data from two utility boilers
operating with correspond!*ng burner designs and coal types. A burner zone
heat liberation rate parameter was used to compare the relative performance
of the different burners under the various firing conditions.
Limited additional testing was conducted to evaluate SO? removal
performance by injected sorbent materials for the different burner designs
and firing conditions. Limestone, hydrated lime, and dolomitic pressure-
hydrated sorbents were injected through various burner passages, and at
various elevations above the burners. Results indicate a strong sensitivity
to injection temperature and the furnace thermal profile.
The work described in this report has been supported by the United
States Environmental Protection Agency through Contract No. 68-02-3130 to
Babcock and Wilcox (B&W) and to Energy and Environmental Research Corporation
under B&W subcontract No. 940962NR.
77^ |
| 11 |

-------
TABLE OF CONTENTS
Section	* Page
1.0 SUMMARY ................ 	 .....	1-1
1.1	Program Pisn » . * * . . . ¦ « * * . . . . . . . . . . *	1—1
1.2	Fuels and Sorbents 					 .	1-4
1.3	Burner Performance and N0X Emissions ..... 		1-4
1.4	SO2 Reduction Potential 			1-7
2.0 INTRODUCTION ... 	 ........	2-1
2.1	DMB Concept and Development 	 .....	2-2
2.2	Program Objectives and History 					2-7
2.3	Guide to the Report . 				2-10
3.0 DMB EVALUATION METHODOLOGY				3-1
3.1	Large Watertube Simulator Tests 		3-3
3.2	Utility Boiler Field Tests 	 .......	3-6
4.0 EXPERIMENTAL SYSTEMS ....... 	 		4-1
4.1	Burner Designs 					4-2
4.2	Fuels and Sorbents ..... 	 .........	4-26
4.3	Test Facility			4-34
4.4	Test Procedures ........... 	 ....	4-44
5.0 BURNER PERFORMANCE AND N0X EMISSIONS			5-1
5.1	Circular Burner .... 	 ......	5-1
5.2	Dual Register Burner ..................	5-6
5.3	120 x 106 Btu/hr Distributed Mixing Burner	5-10
5.4	60 x 106 Btu/hr DMB in Baseline LWS			5-22
5.5	60 x 10® Btu/hr DMB in Insulated LWS .... 	 .	5-38
5.6	Discussion of Results and Extrapolation to
Full Scale 		5-52
v 1

-------
TABLE OF CONTENTS (Concluded)
Sectlon	Page
6.0 S02 REDUCTION POTENTIAL WITH SORBENT INJECTION 		6-1
6.1	Injection Configurations 		6-1
6.2	Test Results			6-5
6.3	Discussion of SO2 Removal Performance ..... 		6-21
7.0 REFERENCES							7-1

-------
LIST OF FIGURES
Figure	Page
1-1	Estimated relationship between injection temperature
and SO2 capture—60 x 10® Btu/hr DMB, Illinois coal .....	1-9
2-1	DMB concept .........................	2—3
2-2	Development history of the Distributed Mixing Burner .... 2-5
3-1	Revised program organization 	 .....	3-2
3-2	N0X emissions vs. boiler load 			 . 3-9
4-1	B&W Circular Burner 					 .	4-3
4-2 Swirl characteristics of 120 x 10® Btu/hr Circular
Burner register 			4-5
4-3 Velocity characteristics of 120 x 10® Btu/hr
Circular Burner 							4-6
4-4 B&W Dual Register Burner . 	 ............	4-7
4-5 Swirl characteristics of 60 x 10® Btu/hr dual register
burner ........ 	 .............	4-9
4-6 60 x 10® Btu/hr Dual Register Burner velocity
characteristics .......... 	 .....	4-11
4-7 Initial 60 x 10® Btu/hr DMB configurations .........	4-12
4-8 Mounting arrangement for 60 x 10® Btu/hr DMBs		 .	4-13
4-9 , Velocity characteristics of initial 60 x 10® Btu/hr
DMB under staged conditions (SRg = 0.70) 		4-15
4-10 Modified 60 x 10® Btu/hr Distributed Mixing Burner 		4-16
4-11 Velocity characteristics of modified 60 x 10® Btu/hr DMB . .	4-17
4-12 Swirl characteristics of 60 x 10® Btu/hr DMB	4-18
4-13 120 x 10® Btu/hr Distributed Mixing Burner with standard
B&W coal impeller	4-20
4-14 Mounting arrangement for 120 x 10® Btu/hr DMB ........	4-21
4-15 Velocity characteristics of 120 x 10® Btu/hr DMB			4-22
1

-------
LIST OF FIGURES (Continued)
Figure , ,	Page
4-16 Swirl characteristics of 120 x 1Q6 Btu/hr DMB			4-23
4-17 Impeller configurations tested with 120 x 10® Btu/hr DMB , .	4-24
4-18 Swirler designs tested with 120 x 10® Btu/hr DMB	4-25
4-19 Coal particle size distributions 			4-32
4-20 Particle size distributions of sorbent materials ......	4-36
4-21 General configuration of Large Watertube Simulator LWS ...	4-37
4-22	Large Watertube Boiler Simulator facility	 .....	4-40
4-23 Insulation pattern in the LWS for operation at
60 x 106 Btu/hr 				4-42
4-24 Flue gas sample train schematic	4-46
4-25	High volume isokinetic particulate sampling system .....	4-49
5-1 Summary of 120 x 10® Btu/hr Circular Burner with Utah Coal .	5-2
5-2 Summary of Circular Burner performance with Illinois coal . .	5-4
5-3 Summary of Circular Burner performance with Comanche coal . .	5-5
5-4 Effect of OFA on Circular Burner performance			5-7
5-5 Effect of excess air on 60 x 10® Btu/hr Dual Register
Burner performance at full load 			 .	5-9
5-6 Effect of excess air on 60 x 10® Btu/hr Dual Register
Burner performance at 75 percent of full load ........	5-11
5-7 Effect of excess air on Phase V DRB configurations'
performance ................. 		5-12
5-8	Summary of 1120 x 10® Btu/hr DMB unstaged operation 		5-15
5-9 Effect of staging on 120 x 10® Btu/hr DMB with Utah coal . .	5-16
5-10	Effect of excess air on 120 x 10® Btu/hr DMB with Utah coal *.	5-17
5-11 Effect of staging on 120 x 10® Btu/hr DMB with alternate
host fuels	5-18
I v i i i ^

-------
LIST OF FIGURES (Continued)
Figure	Page
5-12 Effect of excess air on 120 x 10® Btu/hr DMB staged
operation with alternate host fuels 			 5-12
5-13 Effect of firing rate on DMB 120 x 10® Btu/hr DMB unstaged
operation . . . 					5-21
5-14 Effect of staging on 120 x 10® Btu/hr DMB at reduced load » . 5-23
5-15 Effect of excess air on 120 x 106 Btu/hr DMB at reduced
firing rate				 5-24
5-16 Effect of excess air on unstaged 60 x 10® Btu/hr DMB
performance ......................... 5-26
5-17 Effect of inner register on 60 x 10® Btu/hr DMB staged
performance 	 . 	 ....... 5-27
5-18 Effect of outer register on 60 x 10® Btu/hr DMB staged
performance ......... 	• » * 5-28
5-19	Effect of staging on	60 x 10® Btu/hr DMB performance ....	5-29
5-20	Effect of excess air	on 60 x 10® Btu/hr DMB performance . . .	5-30
5-21	Effect of staging on	60 x 10® Btu/hr DMB with Comanche coal .	5-32
5-22	Effect of staging on 60 x 10® Btu/hr with Illinois coal . . .	5-33
5-23 Summary of effect of excess air on 60 x 10® Btu/hr DMB
staged performance 	 5-35
5-24 Effect of register adjustment on diffuser equipped DMB
performance .... 	 ...... 	 5-37
5-25 Effect of staging on diffuser equipped DMB performance . . . 5-39
5-26 Effect of excess air on diffuser equipped DMB performance . . 5-40
5-27 Effect of staging on 60 x 10® Btu/hr DMB performance in
insulated LWS				 			 5-43
5-28 Effect of excess air oh 60 x 10® Btu/hr DMB performance in
insulated LWS .... 			5-44
5-29 Effect of staging on 60 x 10® Btu/hr DMB with alternate
fuels and insulated LWS	5-45
ix- T	!

-------
LIST OF FIGURES (Continued)
Figure	Page
5-30 Effect of excess air on 60 x 10® Btu/hr DMB with
alternate fuels and insulated LWS 		5-47
5-31	Summary of 60 x 10® Btu/hr DMB burner performance with
Wyodak coal		5-48
5-32 Effect of staging on modified 60 x 10® Btu/hr DMB
performance 		5-51
5-33	Effect of excess air on modified 60 x 106 Btu/hr DMB
performance under staged conditions .............	5-53
5-34 Effect of excess air on modified 60 x 106 Btu/hr DMB
performance 	 . 			5-54
5-35 Effects of burner design and coal type on N0X emissions . . .	5-58
5-36 Correlation of boiler N0X emissions 		5-61
5-37 Correlation of experimental furnace data ..........	5-63
5-38 Effect of thermal environment on N0X emissions from
initial short-flame 60 x 106 Btu/hr DMB			5-66
5-39 Correlation of DMB N0X emissions	5-67
5-40 Effect of burner scale on N0X emission 	 .....	5-69
5-41 Correlation of N0X emissions for the Circular Burner ....	5-71
5-42 Correlation of N0X emissions for the DRB	5-73
6-1	Schematic representation of sorbent injection system ....	6-2
6-2 In-furnace sorbent injection locations ...........	6-3
i
6-3 Summary of sorbent injection with 60 x 10® Btu/hr DMB	,
Illinois coal . 				6-6
6-4	Summary of sorbent injection with 60 x 10® Btu/hr DMB
burner					6-7
6-5 SO2 reduction potential with sorbent injection of the
60 x 10® Btu/hr DMB in the insulated LWS			6-9 j

-------
LIST OF FIGURES (Concluded)
Figure	Page
6-6 Summary of SO2 reduction potential with 60 x 10® Btu/hr
DMB burner with Illinois coal	6-10
6-7 Summary of sorbent injection with 60 x 10® Btu/hr burner
with Utah coal	.	6-12
6-8 Effect of sorbent injection on SO2 reduction potential
with Wyodak coal 	6-13
6-9 Summary of sorbent injection with 120 x 10® Btu/hr
Circular Burner with Utah coal 	6-14
6-10 Summary of sorbent injection with 120 x 10® Btu/hr
Circular Burner with Illinois coal 	 ........ 6-16
6-11 Summary of sorbent injection with the 120 x 10® Btu/hr
DMB firing Utah coal 	 ........... 6-17
6-12 Summary of sorbent injection with staged 120 x 10® Btu/hr
at design firing rate			6-19
6-13 Summary of sorbent injection with staged 120 x 10® Btu/hr
DMB at reduced firing rate	6-20
6-14 Bench scale data comparing SO2 removal performance
for different sorbents			6-24
6-15 Estimated relationship between injection temperature and
SO2 capture—60 x 10® Btu/hr DMB, Illinois coal ....... 6-26
6-16 Small-scale data showing the effect of thermal condi-
tions on SO2 capture			6-27
XI
7^

-------
LIST OF TABLES
Table
1-2
3-1
3-2
4-1
4-2
4-3
4-4
4-5
4-6
4-7
4-8
4-9
4-10
4-11
5-1
5-2
5-3
6-1
6-2
Composition of Test Coals
Comparison of Burner- Performance in the LWS Firing
Utah Coal (SRT = 1.20) 					 . . ,
Boiler Descriptions 		. 		
Burner Operating Characteristics 	
Composition of Test Coals . 		. . . .
Coal Ash Characteristics ........ 	 ,
Predictions of N0X Emissions Based on Coal Composition
• « •
Mean Coal Composition Data—Averages of Ultimate Analysis
Performed on Daily Coal Samples .... 	
Variation of Coal Fineness—Weight Percent Passing 200
Mesh Screen 					
Physical and Chemical Properties of Sorbents
LWS Furnace Characteristics .... 	
LWS Furnace Exit Gas Temperatures 	
Standard Input/Output Measurements 	
Gas Phase Spec.ies Instrumentation	
Detailed Measurement Format 	 . .
t • • •
Summary of the 60 x 10^ Btu/hr DMB Short Flame Burner
Optimum Conditions . 				
Summary of LWS Furnace Temperatures During 60 x 10® Btu/hr
DMB Tests			
Comparison of Burner Performance in the LWS Firing
Utah Coal (SRy = 1.20) . 	
Sorbent Injection Locations Evaluated for Each Burner
Configuration ........... 	 . 	
Summary of Percentage SO2 Removal Data—Illinois Coal,
Injected Ca/S Molar Ratio =2	. ... ,
Page
1-5
1-8
3-1
3-8
4-27
4-28
4-29
4-31
4-33
4-35
4-38
4-43
4-45
4-47
4-50
5-36
5-42
5-57
6-4
6-22
XI i

-------
1.0	SUMMARY
The objective of this program was to demonstrate the performance of the
Distributed Mixing Burner (DMB) on a multi-burner utility boiler. This
involved integrating the DMB concept with Babcock & Wilcox {BAW) burner
components to produce a prototype burner meeting commercial standards. In
the original program plan, the demonstration was to include a full-scale
utility boiler retrofit with Distributed Mixing Burners. The effectiveness
of the DMB was to be determined by direct comparison with the original
equipment burners in one representati ve operating utility boiler. Diffi-
culties in finding a host boiler to participate in a demonstration retro-
fitting existing burners with the new DMB technology resulted in delays to
the overall program. These delays, in turn, caused escalating costs for a
utility boiler retrofit with DMBs. Because of these problems, the program
was restructured to achieve its objective without installation of the DMB in
a utility boiler. The approach taken was extensive testing of DMBs at two
scales and two B&W commercial burner designs in the EPA Large Watertube
Simulator (LWS) coupled with field tests at utility boilers equipped with the
two B&W commercial burners. This approach provided data for burner scaleup,
performance characteristics of the DMB compared to commercial burners, and
commercial burner performance in utility boilers. With this data, the
expected performance of DMBs can be extrapolated to utility boilers with some
confidence.
1.1	Program Plan
In the original program plan, differences in performance with the DMB
were to be determined by direct comparison of the original equipment burners.
The elimination of the field installation precluded this comparison and
required dependence on research furnace test results. As part of the revised
program, the performance of the prototype DMB in the LWS research facility
had to be demonstrated to be similar to the performance in a field operating
boiler. This objective was achieved by: (1) translating developmental DMB
design criteria into practical prototype burners: (2) verifying and
optimizing the performance of the prototype B&W DMBs in the LWS; {3)

-------
evaluating the performance of two commercial burners in both utility boilers
and the LWS; and (4) from that data base extrapolating the prototype DMB
performance to operating utility boilers.
Four different burners were tested:
•	120 x 1Q6 Btu/hr Circular Burner.
•	60 x 106 Btu/hr Dual Register Burner
¦ 60 x 10® Btu/hr Distributed Mixing Burner
> 120 x 106 Btu/hr Distributed Mixing Burner
The first two burners listed represent B&W commercial designs currently in
use in utility boilers. The Circular burner, the B&W pre-NSPS design, was
tested at full scale. The test matrix and measurements bracketed those used
in the field test so that the LWS and field burner performance may be
directly compared. This allowed direct evaluation of furnace environment
effects.
The DRB is the current commercial low-NOx burner design offered by B&W.
Since a full-scale 120 x 106 Btu/hr DRB would be expected to produce flames
about 30-35 ft long and the LWS firing depth is 22 ft, the DRB was tested at
reduced scale, 60 x 10®. B&W estimated that a one-half scale DRB would
produce a flame short enough to avoid flame impingement in the LWS. Reducing
the firing rate by a factor of two from full-scale also reduced the heat
release per unit cooled surface area by a factor of two. This reduced scale
DRB was also tested with additional insulation added to the LWS to more
closely match the thermal environment at full load. This provided a direct
evaluation of the effects of thermal environment independent of burner
scaling. The tests of the DRB similar to the Circular burner tests were
conducted to evaluate burner performance in the LWS.
The full-scale, 120 x 10® Btu/hr Distributed Mixing Burner was the key
to the evaluation of the DMB concept. The test furnace imposed severe
constraints 
-------
which rely on controlled, delayed mixing of fuel with air, generally produce
longer flames than conventional burners. In fact, to produce a flame less
than the furnace depth under staged conditions required more than optimiza-
tion of available burner controls. Iterative modifications were made to
selected burner components, primarily the coal injector, to achieve accept-
able flame dimensions.
Three configurations of the half-scale, 60 x 106 Btu/hr prototype DMB
were evaluated. The two initial configurations considered coal injectors to
produce short vs. long flames. The short flame DMB incorporated a coal
impeller at the end of the coal pipe similar to that used in the Circular
burner. This impeller induces good mixing, producing a relatively short
flame. This DMB design would be appropriate for pre-NSPS boilers with
restricted firing depths. The long flame DMB used a coal diffuser device
like that used in the Dual Register Burner located well back of the burner
exit. It functions to produce a uniform distribution of coal at the exit and
would be expected to result in a long flame similar to that from a DRB.
Based on the developmental tests, long flame DMBs can be optimized to produce
somewhat lower N0X than short flame DMBs. The long flame design would
probably be suitable for retrofit in post-NSPS 8&W units equipped with DRBs
and new boilers with increased firing depth.
During analysis of data from tests with the initial half-scale DMB, the
outer secondary passage was determined to be improperly designed resulting in
unusually high velocities. Following optimization of the full-sized DMB, the
half-scale (60 x 10® Btu/hr) burner was modified to match its design param-
eters. This provided data to determine the effect of burner scaling on
performance to assist interpretation and extrapolation of DMB performance to
utility boilers.
To further aid extrapol ation of the LWS test results, four different
fuels were used. The key fuels were obtained from suppliers of the two host
boilers. Data for each burner were obtained for different fuels with, the
primary objective being to directly link the two host sites, eliminating
questions of fuel composition on scaling.
1-3
-------
1.2
Fuels arid Sorbents
Four different coals were used for the purpose of this test program:
Utah coal, Illinois coal, (high in sulfur), and one coal from each host
boiler site. Data from tests with a fifth coal, Pittsburgh #8, are also
included to broaden the interpretation of results, through the link to the
LIMB/2nd generation burner program. The compositions of these coals are
listed in Table 1-1. The Utah coal has been used as the base fuel at EER in
the development of low-emission, high-efficiency burners. It is a high-
volatile B bituminous coal from the Western United States with a low sulfur
content. The high-sulfur coal used is an Illinois #6 coal. This is a high-
volatile C bituminous coal selected to provide data which would be applicable
to eastern U.S. boilers burning high-sulfur fuels. The Illinois coal has
been tested at EER during previous studies in the LWS, and thus, will permit
comparisons with this program. The Wyodak coal is from the DRB host site,
Wyodak Plant. This is a subbituminous B coal from Wyoming.' Testing the DRB
with the Wyodak coal will establish a link to data from an operating boiler.
Similarly, testing the coal from the Circular burner host site, Comanche Unit
2, provides a second link to operating boiler data.
Three sorbents were used during this program to evaluate the potential
of SO2 reduction with in-furnace injection. Vicron 45-3 limestone and Col ton
hydra ted lime have been used at EER as examples of each type of material in
the development of LIMB technology. In addition, a third sorbent was evalu-
ated for general interest because of its highly reactive nature—pressure
hydrated dolomitic lime. Vicron 45-3 is nominally 99 percent pure CaC03 with
a mass median diameter of 9.8 ^m. The Col ton hydrated lime is nominally 96
percent Ca(OH);? with a median particle size of 4.0ptm. The pressure hydrated
dolomitic lime is a much finer material, with a mass median diameter of only
1.4|xm, in addition to containing a significant amount of magnesium oxide.
1.3	Burner Performance and N0X Emissions
Two commercial B&W designs, the pre-NSPS Circular burner and the low-NOx
Dual Register Burner, were tested in the LWS to provide a basis with which to
1-4
-------
TABLE 1-1. .COMPOSITION OF TEST COALS
Coal
Utah
Illinois
Wyodak
Comanche
Pittsburgh #8
Reporti ng
Basis
As
Rec 'd
Dry
As
Rec 'd
Dry
As
Rec'd
Dry
As
Rec'd
Dry
As
Rec'd
Dry
Proximate
{wt. 1)










Moisture
Ash
Volatile
Fixed C
6.11
8.02
41.26
44.60
0.00
8.55
43.96
46.73
15.26
8.09
34.60
42.06
0.00
9.54
40.84
49.64
23.85
7.17
33.70
35.29
0.00
9.41
44.22
46.37
22.44
5.00
36.12
37,72
0.00
6.45
44,87
48.68
3.50
12.92
33.75
49,83
0.00
13.40
34.98
51.62
Heating
Value










Btu/lb
MMF Btu/lb
MAF Btu/lb
12,288
13,088
14,440
14,311
10,710
12,638
14,209
14,088
8,945
11,753
13,085
12,963
9,325
12,026
12,939
12,855
12,177
12,618
14,876
14,626
III ft mate
(wt. %)










Hoi sture
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen*
6.11
68.58
5.16
1.28
0.60
8.02
10.24
0.00
71.86
5.49
. 1.36
0.64
8.55
10.91
15.26
59.45
4.28
1.07
3.23
8.09
8.64
0.00
70.14
5.05
1.27
3.81
9.54
10.21
23.85
50.93
3.65
0.75
0.43
7.17
13.23
0.00
66.89
4.81
0.98
0.57
9.41
17.34
22.44
54.25
3.80
0.76
0.43
5.00
13.32
0.00
69.97
4.91
0.98
0.56
6.45
17.14
3.50
68.13
4.63
1.21
3.22
12.42
6.41
0.00
70.54
4.79
1.26
3.30
13.40
6.63
Forms of
Sulfur
(wt. %)










Sulfate
Puritic
Organic
0.01
0.13
0.46
0.01
0.13
0.50
0.18
0.95
2.11
0.21
1.11
2.49
0.01
0.06
0.36
0.01
0.09
0.47
0.02
0.09
0.32
0.02
0.12
0.42
0.22
1.62
1.38
0.23
1.65
1.42
*Oxygen determined by difference.
1-5
-------
judge DMB performance. This comparative evaluation verified safe, efficient
operation of the prototype DMB providing confidence for field application.
Limited sorbent injection tests evaluated the effect of burner design on SO2
reduction potential for both near burner and upper furnace locations.
The full-scale 120 x 106 Btu/hr Distributed Mixing Burner was the key to
this demonstration program. The LWS test furnace imposed severe constraints
to flame shape and size for a 1 ow-NOx burner. Low-N0x burners, like the DMB,
rely on controlled, delayed mixing of the fuel with air. This delayed mixing
generally produces a long flame which may cause operational problems in a
boiler. Although equipped with adjustable inner and outer secondary air
control as well as the tertiary air ports, the dominant factor in determining
ultimate performance (N0x/f1ame length) was the coal injector configuration.
Iterative modifications were made to the coal injector to yield the optimum
performance for the LWS. There was a direct tradeoff between N0X emissions
and flame length.
The final design- selected resulted in unstaged flames about 16 ft long.
Under staged conditions, with a burner zone stoichiometric ratio (SRb) of
0.70, the flame length increased to approximately 22 ft. The optimum
configuration for the 120 x 106 Btu/hr DMB was determined to be:
•	Spreader design = 4-inch support pipe with four 8-inch blades at a
300 angle from axial.
•	Inner Spin Vanes = 35° Open CW.
9 Outer Register = 10° Open CW (clockwise).
•	Spreader = 3 - i n . retracted inner/outer secondary air
distribution--50/50.
N0X emissions for the DMB at these optimum settings at nominal full-load
conditions with a burner zone stoichiometry of 0.70 and 20 percent excess air
were 300, 340, 298, and 273 ppm (corrected to 0% O2) for Utah, Illinois,
1-6
-------
Comanche, and Wyodak coals, respectively. This performance compares
favorably with the two commercial B&W burners tested, as seen in Table 1-2.
1.4 ' SO2 Reduction Potential
The potential for SO2 control combined with N0X reduction was evaluated
in a series of sorbent injection trials. A total of six different injection
locations were considered. Three sorbents were used: Vicron 45-3 limestone,
Col ton hydrated lime, and a limited number of tests with a pressure hydrated
dolomitic lime. Thermal environment was the key factor determining SO2
capture efficiency. The sensitivity of SO2 capture to thermal environment is
summarized in Figure 1-1. Upper furnace locations where gas temperatures
were about 2200°F yielded the highest captures. Near burner injection,
either with the coal or through tertiary air ports generally gave the poorest
SO2 capture. The pressure hydrated dol omitic lime was the most effective of
three sorbents on a Ca/S molar ratio basis, however, the advantage disappears
when considered on a mass basis because of the additional magnesium
component.
1-7
-------
TABLE 1-2;. .COMPARISON OF BURNER PERFORMANCE IN. THE
LWS FIRING UTAH COAL (SRT = 1.20)
j	'	. O; '	1	' ' ) }

DMB
DRB
Circular-

Full Scale
Half Scale
Half Scale
Full Scale
Firing Rate (106 Btu/hr)
-' 120
60
60 "
120
' srb
0,70
0.70
1.20
1.20
FEGT (OF)
1792
1776
1776
1828
N0X {ppm 0 0« O2)
.{|00>
350
390
380
Flame Length (ft)
22
18
18
>22
/ •' .
• SR_ = Burner Zone Stoichiometric Ratio
' D 1 ¦ ?
FEGT = Furnace Exit Gas Temperature
\
K'
' ^
"
¥ : ^ J-*.-
i . ' >
-------
60

CD
s*.
CL
CO
<->
CVJ
o
CO
40
20
CaCOg
A Ca(0H)2
Solid Symbols--Insulated Furnace
Open Symbols--Baseline Furnace
1
1
_L
_L
1800	2000	2200	2400
Estimated Temperature at Injection Plane (°F)
'Figure 1-1'.
Estimated relationship between injection temperature and
S09 capture--60 x 10*> Btu/hr DMB, Illinois coal.
-------
2.0	INTRODUCTION
N0X emissions from pulverized coal combustion are typically higher than
from the combustion of liquid or gaseous fossil fuels because coal contains
substantial quantities of nitrogen compounds. During combustion, these
compounds decompose to liberate HCN and NH3 which react readily with oxygen
to form N0X. Up to 80 percent of the total N0X emissions from the combustion
of coal is due to fuel nitrogen oxidation.!
One of the most effective techniques for reducing N0X emissions from
high-nitrogen fuels is staged combustion. This involves firing the fuel
under oxygen deficient conditions initially, followed by secondary air
addition to complete combustion. In the fuel-rich primary zone, the bound
nitrogen compounds are preferentially reduced to N2 prior to the addition of
the secondary air. The effectiveness of this type of staging in reducing M0X
emissions depends on the combustion conditions, particularly in the initial
fuel-rich zone. The optimum stoichiometry in the fuel-rich zone for N0X
abatement is about 0.7, or 70 percent of theoretical air required for
stoichiometric combustion. Staging has been demonstrated to be effective in
reducing N0X emissions on full scale wall-fired boilers through the use of
overfire air ports.2 However, in boiler retrofit applications of overfire
air for staging, the stoichiometry in the initial zone around the burners is
preferably maintained above stoichiometric conditions (100 percent
theoretical air) to minimize slagging and corrosion in the lower furnace and
to achieve acceptable char burnout. This limits the effectiveness of
overfire air ports for N0X control.
For several years, Energy and Environmental Research Corporation (EER)
has been working with the U.S. Environmental Protection agency (EPA) in the
development of a 1ow-N0x pulverized coal burner for wall-fired applications.
The Distributed Mixing Burner (QMB) approach involves staging the combustion
process with discrete air ports around each circular burner. The DMB allows
a fuel-rich primary zone, with stoichiometry near the optimum range, to be
established adjacent to the burner while maintaining an overall oxidizing
atmosphere in the furnace around the burners, minimizing slagging, corrosion,
2-1
-------
and char burnout problems. In the development process, N0X emissions less
than 0.15 lb/lO^ Btu were achieved with, test burners in research facilities.3
The objective of this program was to demonstrate the performance of the
DMB on a multi-burner utility boiler. This involved integrating the DMB
concept with Babcock & Wilcox (B&W) burner components to produce a burner
meeting commercial standards. Performance of such a prototype B&W DMB was to
be verified in the EPA Large Watertube Simulator (LWS) test facility,
followed by full retrofit and testing in a utility boiler.
2.1	DMB Concept and Development
The DMB concept involves staging the combustion process to minimize N0X
emissions while maintaining an overall oxidizing atmosphere in the furnace to.
minimize slagging and corrosion. N0X production from fuel nitrogen compounds
is minimized by driving a majority of the compounds into the gas phase under
fuel-rich conditions and providing a stoichiometry/time-temperature history
which maximizes the decay of the evolved nitrogen compounds to N2- Thermal
M0X production is minimized by heat loss from- the fuel-rich zone which
reduces peak temperatures.
A schematic representation illustrating how the DMB design stages the
fuel/air mixing sequentially is shown in Figure 2-1. The combustion process
occurs in three zones. In the first zone pulverized coal transported by the
primary air combines with the inner secondary air to form a very fuel-rich
(30 to 50 percent theoretical air) recirculation zone which provides flame
stability. The coal devolatizes and fuel nitrogen compounds are released to
the gas phase. Outer secondary air is added in the second "burner zone"
where the stoichiometry increase up to about 70 percent theoretical air.
This is the optimum range for reduction of bound nitrogen compounds to N2«
Air to complete the combustion processes is supplied through tertiary ports
located outside the burner throat. This allows substantial residence time in
the burner zone for decay of bound nitrogen compounds to N2 and radiative
heat transfer to reduce peak temperatures. The tertiary ports surround the
2-2
-------
TERTIARY AIR
INNER
SECONDARY AIR
X
COAL AND_,
PRIMARY AIR
OUTER
SECONDARY AIR
X
VERY FUEL RICH PROGRESSIVE AIR ADDITION ZONE
ZONE (AVERAGE (OVERALL STOICIIIOMETRY 70%)
STOICHIOMETRY 40%)
FINAL AIR ADDITION ZONE FOR BURNOUT
(OVERALL STOICIIIOMETRY 120%)
Figure12-iT'. DMB ;c o n c ep t
-------
burner throat providing an overall oxidizing atmosphere and minimize
interactions between adjacent burners.
Components for a typical, fully commercial DMB design would include:
•	Four independently controlled air streams:
Primary air for pneumatic transport of coal from pulverizer to
burner.
Two concentric annular secondary air streams around the
primary jet.
Tertiary air through four outboard ports.
•	Fuel injector design to produce a uniform coal distribution and
initiate mixing with secondary combustion air streams to stabilize
the f1ame.
•	Adjustable assemblies for each secondary air stream to control air
flow rate distribution and degree of swirl.
•	Commercial ignition and flame scanner, system for start-up and
safety.
The key components for the DMB are the outboard staged air ports closely
coupled to each burner in a multi-burner installation.
The development history of the DMB is sunmarized in Figure 2-2. Initial f"
development of the DMB concept was carried out at the International Flame
Research Foundation (IFRF) under EPA Contract 68-02-0202. This included
proof-of-concept tests in a research furnace firing at 8.5 x 106 Btu/hr.
Additional development and scale-up tests were conducted at EER under EPA
Contract 68-02-1488. To provide a standard means of evaluating burner
performance, two large-scale test facilities were constructed: the Small
2-4
-------
r>o
i
lti
L & C STEINMULLER
DMB
DEVELOPMENT
LCS SM
BURNER
	I
RILEY STOKER
DMB
DEVELOPMENT
RILEY DMB
LIMB
CONCEPTS
/ METHODOLOGY
AFOR PREDICTING DMB
V PERFORMANCE V
FIELD
DEMONSTRATION
NOTE: NUMBERS ARE
EPA CONTRACT
' GENERIC >
2nd GENERATION
10W NO BURNER
INTEGRATION WITH
BABCOCK & WILCOX
68-02-3130
INTERNALLY STAGED
BURNER DEVELOPMENT
68-02-3923
68-02-2667
ADVANCED
CONCEPTS
INTEGRATION WITH
FOSTER WHEELER
COMPARATIVE TESTS
RILEY STOKER DMB
68-02-3913
68-02-1488
SCALE-UP
TESTS
COMPARATIVE TESTS
LES SM. BURNER
68-02-3916
INITIAL DMB
CONCEPT
68-02-0202
Figure 2-2. Development history of the distributed mixing burner!
-------
Watertube Simulator (SWS) and the Large Watertube Simulator. The SWS was
designed to simulate the thermal environment of a small watertube boiler. It
had a capacity of 10 x 10® Btu/hr and provided a means of "testing burners at
moderate scale so that the effects of parametric variations could be
evaluated at low cost. The LWS was designed to simulate a large industrial
or small utility watertube boiler. The furnace shape was similar to
commercial boilers with a hopper bottom, a nose, and provision for front-wall
firing. The firing capacity of the LWS was 150 x 10® Btu/hr which allowed
full-scale burners to be evaluated under conditions simulating large
commercial systems.
Six experimental DMB configurations were designed and tested. The
designs covered a firing rate range of 10 to 100 x 10® Btu/hr as single
burners and also as a four-burner array. The DMBs were research designs with
flexible parameters so that the effects of burner design variations on N0X
emissions, flame stability and combustion efficiency could be evaluated.
Minimum N0X emissions for each burner were in the range of 0.1 - 0.2 lb/10®
Btu. The results of the tests were compiled into a set of design criteria
which could be used to apply the DMB concept to commercial burners.
At this point, the development divided into two parallel efforts. One
focused on advanced concepts of N0X and S0X control. In EPA Contract
68-02-2667 the DMB concept was integrated with sorbent injection for S0X
control. The resulting process was termed "Limestone Injected Multi-Stage
Burner" (LIMB). In EPA Contract 68-02-3923 advanced burner concepts were
developed to achieve staging without the need for the outboard tertiary air
ports of the DMB.
The other effort focused on the further development and commercializa-
tion of the DMB. Two EPA demonstration programs were established. This
project, EPA Contract 68-02-3130, was to demonstrate the application of DMB
technology to large utility boilers with B&W burner hardware. In a parallel
project (EPA Contract 68-02-3127) the DMB concept was integrated with Foster
Wheeler (FW) components to be demonstrated in a small utility boiler in the
range of 100 to 500 x 103 lb/hr.
2-6
-------
In parallel with this EPA development, two other burner manufacturers
elected to develop DMBs for commercial offerings. Based on the results of
the initial IFRF tests, L&C Steinmuller {LCS), a German burner/boi1er
manufacturer, developed the Staged Mixing (SM) burner based on the DMB
concept, LCS installed the SM burners in a 700 MW German boiler and
demonstrated a 50 percent reduction in N0X emissions.4 The SM burner is now
the standard commercial burner offering for LCS. Riley Stoker developed a
DMB based on the design criteria developed in EPA Contract 68-02-1488 in
conjunction with EER. The Riley Stoker DMB was tested in research facilities
at EER. The results of these independent developments were integrated with
the EPA work in EPA Contracts 68-02-3916 and 68-02-3913 which included tests
of the LCS and Riley Stoker DMBs, respectively, in the LWS test facility.
2.2	Program Objectives and History
The objective of this program was to evaluate the performance of the EPA
Distributed Mixing Burner, incorporating B&W burner hardware, in a utility
boiler. The original plan to achieve this objective after contract
initiation on September 30, 1978 involved four key elements:
1.	A field test of the host boiler with the original burners to
establish the "baseline" burner/boiler performance.
2.	A test of the original, "baseline" burner in the LWS research
furnace to calibrate the furnace against the corresponding host
boiler.
3.	Evaluation and optimization of a prototype DMB with B&W components
in the LWS to verify performance prior to installation at the host
site,.
4.	Long-term field evaluation of the DMB in the host boiler.
-------
Babcock & Wilcox held the prime contract with the EPA to achieve these
goals- EER was subcontracted to support the B&W effort. The EER effort
included;
• Engineering assistance throughout the program as required, such as
DMB design input, definition of program plan and measurement plan,
data analysis, and reporting,
« All LWS testing of the original baseline burner and the prototype
DMB.
¦ Field testing support during both the baseline and low-NOx test
phases.
Progress on this plan was delayed because of difficulties in finding a
suitable host boiler. The selection of the host site was the key to the
entire project. The host site defined boiler specific burner design
requi rements, including firing capacity, burner and tertiary air port
spacing, fuel characteristics, nominal operating conditions and duty cycle,
furnace dimensions, and flame confinement. In about May 1981, a final effort
to secure a host failed. A proposed Ohio Edison boiler was found to be
unacceptable by the EPA due to high projected cost to complete this project.
Costs to fully retrofit a uti1ity boiler with Distributed Mixing Burners had
escalated beyond available funding. From that point, the effort focused on
restructuring the program to achieve the program goals without a costly field
demonstration. Negotiations among the participants, B&W, EER, and EPA,
reached initial agreement on a revised program about December 1981. The
revised program scope was finalized in February 1983, with the cost breakdown
agreed to in November 1983. The revised program scope addressed two distinct
issues. The major portion of the program still focused on the evaluation of
the Distributed Mixing Burner for utility boiler application. The second
area of interest added to the scope of this project was an evaluation of
alternate concepts for 1 ow-NOx emissions coupled with high level,s of
particulate removal and possible SQX control. These alternate concepts,
detailed in Volume IV of this report, considered fuel-rich, high-temperature
1 2-8
-------
prechambers, such as cyclone furnaces. This alternate concept program was
structured to: (1) compile and synthesize existing data on coal-fired
precombustion systems; (2) conduct initial pilot scale tests at 1 x 106 But/
hr to identify the key parameters affecting N0X and S0X reduction potential,
and (3) a second phase of more fundamental testing structured to investigate
a broader range of S0X control issues in smaller, well-controlled experiments
to generate a more complete set of basic precombustor design data.
The evaluation of the DMB for utility boiler application was
restructured to achieve the program objective without a field installation.
In the original program plan, differences in performance with the DMB were to
be determined by direct comparison of the or1 ginal equipment burners. The
elimination of the field installation precluded this comparison and required
dependence on research furnace test results. As part of the revised program,
the performance of the prototype DMB in the LWS research facility had to be
demonstrated to be similar to the performance in a field operating boiler.
This objective was achieved by: (1) translating developmental DMB design
criteria into practical prototype burners; (2) verifying and optimizing the
performance of the prototype B&W DMBs in the LWS; (3) evaluating the
performance of two commercial burners in both utility boilers and the LWS;
and (4) from that data base extrapolating the prototype DMB performance to
operating utility boilers.
Work began in earnest on the revised program scope for the DMB
evaluation in May 1984 with the preparation of a detailed test plan. Actual
testing for this program was initiated in September 1984 and was completed in
June 1986. During that time frame, the scope of the DMB evaluation program
was expanded further. The expanded scope was an opportunity presented by the
initiation of another EPA demonstration project to directly participate in
full-scale application of low-emission burners to an operating boiler.
The EPA demonstration of LIMB (Limestone Injection Multistage Burner)
technology had the objective of reducing both N0X and SO2 emissions by 50
percent. The N0X reduction was to be achieved by retrofitting existing
burners at Ohio Edison's Edgewater Unit 4 boiler with second generation low-
2-9
-------
N0X burners. Because of the constraints at this boiler, evaluation of three
candidate B&W burners prior to selection was essential.
The three B&W 1ow-NGx burner designs considered, the Dual Register
Burner (DRB), Babcock-Hitachi N0X Reducing (HNR) burner, and the XCL burner,
were tested at full scale in the EPA Large Watertube Simulator (LWS) to
determine the optimum design for use at the Edgewater boiler as part of this
contract. Burners sized at 78 x 106 Btu/hr, the same size as the Edgewater
burners, were tested in the LWS, minimizing scale-up questions. By
coincidence, the LWS has a firing depth of 22 ft, essentially the same as
Edgewater Unit 4. Screening tests of the three basic burner designs were
conducted firing Pittsburgh #8 coal, the coal to be used during the LIMB
demonstration, to determine optimum operating conditions. In these tests,,
the influence of adjustable burner parameters (e.g. swirl level, air
distribution) and of changes to burner hardware components (e.g. coal nozzle)
was determined. For each of the three basic burner designs one configuration
was selected for sorbent injection testing, to determine the effect of burner
design on SO2 capture. Following the screening tests of the three burners,
selected XCL burner configurations were characterized with three additional,
distinctly different coals to broaden the applicability of this new burner.
These tests were conducted in two phases, with the initial screening tests
from August through October 1985 and the final optimization tests from
February through March 1986.
2.3	Guide to the Report
The broad scope of this program can be separated into four distinct
parts; (1) the evaluation of prototype DMBs for application to utility
boilers; (2) field tests of baseline burners at two host boilers to support
the extrapolation of prototype DMB performance to field applications; (3)
evaluation of three B&W second generation low-NOx burners to be selected for
use in the EPA LIMB demonstration; and (4) alternate concepts for N0X and S0X
control in precombustors. Each of these represents a distinet element of the
program. This report is, therefore, organized to fully address each element.
! 2-10
-------
Volume I--Distributed Mixing Burner Evaluation. Thts volume of the
report, Part I, presents the results from the prototype burner evaluations in
the LWS, the principal element to achieve the original program objectives.
This part describes the methodology employed to evaluate the DMBs without a
field retrofit, linking research furnace results to operating boilers. The
experimental systems, including test burners, fuels, the test facility
itself, and testing procedures, are fully detailed. Burner performance for
each test burner are discussed. The key to interpretation of the results is
the link of the LWS test results to operating utility boilers achieved with
tests of commercial 8&W burners in the LWS and field test results of the same
burner design in utility boilers. This link allows extrapolation of
prototype DMB performance from the LWS to the field. A summary of sorbent
injection trials for SO2 control is also included in Volume I to broaden the
existing data base and experience with LIMB technology.
Volume II—Second Generation Low-NOx Burners. Volume II summarizes the
LWS trials of the three B&W 1ow-NOx burners considered for the EPA LIMB
demonstration program at Edgewater Station Unit 4. The three burners
included: the Dual Register burner (DRB), Babcock-Hitachi NR burner (HNR),
and the B&W XCL burner. The burners and each configuration tested are
described, along with the fuels and test facility configuration used
throughout these tests. The optimization of the various configurations of
each basic burner design is described with respect to the key performance
criteria of N0X emissions, flame length, combustion efficiency, and burner
pressure drop. The performance of each optimized configuration is compared
to the LIMB demonstration site requirements and recommendations for burner
selection are made. Finally, a brief series of sorbent injection tests was
conducted for a selected configuration of each burner design. These tests
were performed to determine any possible effect of burner design on SO2
capture potential with sorbent injection.
Volume III — Field Evaluations. Volume III details the field tests
performed in conjunction with the DMB evaluation. The field tests were
performed at two different utility boilers, generally similar in design and
size except for the burner equipment. Comanche Unit 2 of Colorado Public
2-11
-------
Service was equipped with B&W Circular burners, the pre-NSPS (New Source
Performance Standard) burner design. The Wyodak Plant of Black Hills Power
was equipped with Dual Register Burners. Test results of emissions and
boiler performance are presented for each unit. Key performance aspects from
these two boilers are used in interpretation of LWS tests of the Circular
burner and Dual Register burner.
Volume IV--A1ternate Concepts. Precombustor studies for N0X and SO2
control are described in Part IV of this report. This work represents
alternate concepts considered as a result of the program's reorganization.
Part IV stresses the fundamental design considerations for precombustor
control of SO2 emissions with a brief summary of pilot scale, 1 x 106 Btu/hr
tests for N0X control. The various experimental apparatus and test
procedures for this fundamental work are described. Results from entrained
flow sulfidation tests and slag sulfur chemistry are fully detailed.
Volume V--Burner Evaluation Data Appendices. Volume V of this report
documents the Quality Assurance program for the LWS tests of the DMB
evaluation and the Second Generation Low NOx burner selection. In addition,
computer listings of all valid data reported in Volumes I and II are included
for reference.
2t-12
-------
3.0 DMB EVALUATION METHODOLOGY
In the original program plan, differences in performance achieved with
the DMB were to be determined by direct comparison of the original equipment
burners and the DMBs in one representative operating utility boiler. The
elimination of the field installation of the DMB precludes this comparison
and requires dependence on research furnace test results. As part of the
program, it must be verified that the performance of burners tested in the
LWS is similar to, or may be extrapolated to, the performance of burners
installed in a uti 1 i ty boiler. Two factors which can influence this
extrapolation are burner scaling and furnace environment. The revised
program plan provided an evaluation of the scaling and furnace environment
effects through two field tests of utilities equipped with conventional B&W
burners and tests of five burners in the EPA LWS at EER. The revised program
organization is shown in Figure 3-1.
Two host boilers were selected. These boilers were hosts in the sense
that they were field tested and the commercial burners tested in the LWS were
designed to match the specific host boiler characteristics. The sites for
the host boilers were:
1.	Wyodak Plant. This unit is a B&W opposed wall-fired boiler
equipped with 30 Dual Register Burners (DRB). The plant is rated
at 330 MWe with a nominal maximum capacity of 350 MWe* The plant
is located near Gillette, Wyoming, and is operated by Pacific Power
and Light Company and Black Hills Power & Light Company.
2.	Comanche Station #2. This unit is equipped with 32 B&W Circular
burners in an opposed wall-fired arrangement. The plant, owned and
operated by Colorado Public Service Company, has a generating
capacity of 350 MWe.
Ideally, the burner tests in the LWS would be conducted at full-scale so
that burner scaling methods need not be considered. However, the flame
length from a full-scale, 120 x 106 Btu/hr B&W DRB was expected to exceed the
3-1
-------
BOILER
SELECTION
CIRCULAR
BURNER HOST
BOILER
FIELD
FIELD TEST
CIRCULAR
DRB
HOST BOILER
FIELD TEST DMB
PROTOTYPE
DMB
DESIGN
Figure 3-1
FULL SCALE
CIRCULAR
h SCALE DRB
h SCALE DRB
INSULATED
FULL SCALE
DMB
h SCALE DMB
INSULATED
OVERALL
FIELD/LWS
COMPARISON
EFFECTS OF
HEAT RELEASE/
SURFACE AREA
OPTIMUM
SHORT FLAME
DMB
OPTIMUM
LONG FLAME
DMB
EVALUATION
OF DMB
CONCEPT
Revised program organization.
-------
firing depth of the LWS. Consequently, the DMB tests were performed at 120 x
10^ Btu/hr and 60 x 10^ Btu/hr in order to better understand scaling effects,
and to tie all aspects of the test program together, i.e. scaling within the
LWS and scaling from LWS to field. Burner scaling, however, was not well
understood at the time.
In previous developmental DMB tests, burner scaling was based on
constant velocities and geometrical similarity. Results from these tests
suggest that this type of scaling results in increasing N0X emissions with
firing rate. From a phenomenological viewpoint, scaling the velocities with
the linear dimensions should be used to maintain constant flame residence
times. Practical considerations, such as maximum acceptable pressure drop
and minimum velocities to maintain pneumatic pulverized coal transport, limit
the application of this scaling method. This program therefore included
tests to confirm that the sealing method employed gives satisfactory results.
It is well known that furnace environment can affect burner performance.
The important variables include furnace geometry, flame interaction, furnace
surface area and volume, and furnace heat extraction. The burner/boiler
manufacturers, including B&W, have found that N0X emissions from burners of
their design can be correlated with the furnace heat release per unit cooled
surface area. The definitions of the cooled surface area and the shapes of
the correlations differ among the manufacturers. The program included tests
to evaluate the effects of the specific LWS furnace environment on burner
performance and to compare these effects with those of full-scale utility
boilers.
3.1	Large Watertube Simulator Tests
The evaluation of the DMB for utility boiler application was
restructured, to achieve the program objective without a field installation.
In the original program plan, differences in performance with the DMB were to
be determined by direct comparison of the original equipment burners. The
elimination of the field installation precluded this comparison and required
dependence on research furnace test results. As part of the revised program,
3-3.
-------
the performance of the prototype DMB in the LWS research facility had to be
demonstrated to be similar to the performance in a field operating boiler.
This objective was achieved by: (1) translating developmental DMB design
criteria into practical prototype burners; (2) verifying and optimizing the
performance of the prototype B&W DMBs in the LWS; (3) evaluating the
performance of two commercial burners in both utility boilers and the LWS;
and (4) from that data base extrapolating the prototype DMB performance to
operating utility boilers.
Four different burners were tested:
•	120 x 10® Btu/hr Circular Burner
•	60 x 10® Btu/hr Dual Register Burner
•	60 x 10® Btu/hr Distributed Mixing Burner
•	120 x 10® Btu/hr Distributed Mixing Burner
The first two burners listed represent B&W commercial designs currently in.
use in utility boilers. Due to the characteristic short flame, the Circular
burner, the B&W pre-NSPS design, was tested at full scale. The test matrix
and measurements bracketed those used in the field test so that the LWS and
field burner performance may be directly compared. This allowed direct
evaluation of furnace environment effects.
The DRB is the current commercial low-NOx burner design offered by B&W.
Since a full-scale 120 x 10® Btu/hr DRB would be expected to produce flames
about 30-35 ft long and the LWS firing depth is 22 ft, the DRB will be tested
at reduced scale, 60 x 10® Btu/hr. B&W estimated that a one-half scale DRB
would produce a flame short enough to avoid flame impingement in the LWS.
Reducing the firing rate by a factor of two from full-scale also reduced the
heat release per unit cooled surface area by a factor of two. Therefore, the
reduced scale DRB was also tested with additional insulation added to the LWS
to more closely match the thermal environment at full load. This provided a
direct evaluation of the effects of thermal environment independent of burner
scaling. The tests of the DRB, similar to the Circular burner tests, were
conducted to evaluate burner performance in the LWS.
3-4
-------
For the full-scale, 120 x 106 Btu/hr Distributed Mixing Burner, (DMB)
the test furnace imposed severe constraints with regard to flame shape and
size. Low-N0x burners, generally produce longer flames than conventional
burners, and the available firing depth in the LWS is considered to be
comparatively small for burners of this capacity. Iterative modifications
were made to selected burner components, primarily the coal injector, were
therefore necessary in order to achieve acceptable flame dimensions with the
120 x 106 Btu/hr DMB.
Three configurations of the half-scale, 60 x 106 Btu/hr prototype DMB
were evaluated. The two initial configurations considered coal injectors to
produce short vs. long flames. The short flame DMB incorporated a coal
impeller at the end of the coal pipe similar to that used in the Circular
burner. This impeller induces good mixing, producing a relatively short
flame. This DMB design would be appropriate for pre-NSPS boilers with
restricted firing depths. The long-flame DMB used a coal diffuser like that
used with the Dual Register Burner, located well back from the burner exit.
It functions to produce a uniform distribution of coal at the exit and would
be expected to result in a long-flame similar to that from a DRB. Based on
the developmental tests, long flame DMBs can be optimized to produce somewhat
lower N0X than short-flame DMBs. The long-flame design would probably be
suitable for retrofit in post-NSPS B&W units equipped with DRBs and new
boilers with increased firing depth.
During analysis of data from tests with the initial half-scale DMBs, the
outer secondary passage was determined to be improperly designed resulting in
unusually high velocities. Following optimization of the full-sized DMB, the
half-scale 60 x 106 Btu/hr burner was modi fi ed to match its design
parameters. This provided data to determine the effect of burner scaling on
performance to assist interpretation and extrapolation of DMB performance to
utility boilers.
To further aid extrapolation of the LWS test results, four different
fuels were used. The key fuels were obtained from suppliers of the two host
boilers. Data for each burner were obtained for different fuels with the
3-5
-------
primary objective Being to directly link the two host sites to tests
performed in the LWS, eliminating questions of fuel composition on scaling.
3.2	Utility Boiler Field Tests
Table 3-1 shows the characteristics of the two boilers tested. Both
boilers fire subbi tuminous coal and use a front and rear wall firing
configuration. The front and rear wall burners at the Comanche boiler are
directly opposed with four rows of four burners each. The front and rear
burners at Wyodak are offset to avoid flame interactions and are arranged in
five rows of three burners each. The boilers have comparable furnace cross-
sectional dimensions, but the Wyodak boiler has a taller furnace to accom-
modate the five burner rows. Thus, the Wyodak furnace has a lower ratio of
firing rate to cooled surface area.
During testing, the boilers were generally operated in a normal fashion
by the operators without interference from EER. Thus, the burner settings,
load, and excess air were controlled by plant personnel. The overfire N0X
ports were closed during the day at the Comanche boiler at the request of
EER, and returned to their normal open position of 18 percent at night.
Table 3-2 shows the typical- burner settings and flame character!sties during
the tests. Both the Circular and Dual Register burners operated satisfac-
torily during the tests. Exact flame lengths could not be determined with
the available observation ports. Both burners showed a high combustion effi-
ciency and large imbalances of fuel or air distribution were not observed.
Both boilers operated over a narrow excess O2 range, 2.5 to 3.5 percent
at Comanche and 3.8 to 4.0 percent at Wyodak. Thus the data were not
sufficient to establish N0X emissions with excess O2. Figure 3-2 shows N0X
emissions at the two boilers as a function of load. Both correlations show a
similar slope, with lower M0X emissions for the Dual Register burner at
Wyodak. Nominal N0X emissions with the Circular burner at Comanche were 550
ppm at OS O2 (0.64 lbs/106 Btu). Full load emissions at Wyodak with all
mills in service were 395 ppm at 0% O2 (0.46 lbs/106 Btu). More detailed
results are presented in Part III.
3-6
-------
TABLE ,3,-1. BOILER DESCRIPTIONS
UNIT ;
COMANCHE UNIT 2
WYODAK UNIT 1
UTILITY
COLORADO PUBLIC SERVICE
PACIFIC POWER & LIGHT
BLACK HILLS POWER & LIGHT
BOILER MANUFACTURER
B&W
B&W
YEAR OF INITIAL OPERATION
1976
1978
GROSS GENERATING CAPACITY
35d MWe
350 MWe
TYPE OF BURNER
B&W CIRCULAR
B&W DUAL REGISTER
NO. OF BURNERS
32
30
NO. OF MILLS
4
5
BURNER ARRANGEMENT
4W x 4H ON
FRONT & REAR WALLS
3W x 5H ON f
FRONT & REAR WALLS
FURNACE DIMENSIONS
43'W x 45'D x 161'H
46'W x 45'D x 180'H
COAL TYPE
SUB BITUMINOUS
SUB BITUMINOUS
ADDITIONAL FEATURES
8 NOx PORTS
SEALED N0X PORTS
OPPOSING BURNERS OFFSET
COMPARTMENTED WINDBOXES
-------
TABLE 3-2. BURNER OPERATING CHARACTERISTICS
V
UNIT
COMANCHE
WYODAK
BURNER TYPE
CIRCULAR
DUAL REGISTER
TYPICAL BURNER
SETTINGS
BOTTOM BURNERS-REGISTERS 50% OPEN
TOP BURNERS -REGISTERS 100% OPEN
OUTER REGISTERS 50% ^OPEN
INNER REGISTERS 25-50% OPEN
SWIRL VANES 10% OPEN
FLAME CHARACTERISTICS
LONG AND NARROW
0-2 FT STANDOFF
LONG AND NARROW
35 - 44 FT l'oNG
0.5-3 FT STANDOFF
-------
o.
Q.
U
-a
800
700
600
500
CM
O 400
o
©
X.
o
300
200
100

MCR = 350 MWg (Gross)
(Wyodak and Comanche)
MCR
\v
•. / <$T4-
y -
5% 1
a
^>v.
5 V
60	120 180 240 300 360
Load MWfi
Fi gure. 3-2. N0X emissions vs. boiler load.
-------
4.0	EXPERIMENTAL SYSTEMS
Four different burners were tested in the EPA Large Watertube Simulator
to facilitate evaluation of the Distributed Mixing Burner for utility boiler
application without a costly retrofit installation. The burners tested
included:
•	120 x 106 Btu/hr B&W Circular Burner
•	60 x 10® Btu/hr B&W Dual Register Burner
•	60 x 10® Btu/hr Distributed Mixing Burner
•	120 x 106 Btu/hr Distributed Mixing Burner
The Circular burner and the Dual Register burner are both B&W commercial
burner designs currently installed in utility boilers. Characterizing the
performance of these two commercial designs in the LWS essentially calibrates
the test furnace, establi shi ng the basis with which to extrapolate the
performance of the prototype Distributed Mixing Burner to operating boilers.
The Distributed Mixing Burner was evaluated at two scales to develop scale-up
criteria for burner performance from the test faci1ity to an operating
boiler.
Evaluation of these test burners was conducted in the EPA Large
Watertube Simulator at the El Toro, California, test site of EER. To
determine effect of furnace thermal environment on performance and resulting
scale-up criteria, two insulation configurations were established for
selected burner trials. Burner performance was evaluated using four
principal coals: a base development fuel, a high sulfur coal for SO2 control
considerations, and two coals from the host boiler sites of the commercial
burners. In addition to burner performance and N0X emission optimization,
SO2 reduction potential was evaluated by injection of three sorbent materials
through several furnace locations. All tests were conducted in the LWS in
accordance with established Quality Assurance procedures following EPA
guidelines. Documentation of the Quality Assurance program is in Part V,
Appendix A of this report.
4-1
-------
4.1
Burner Designs
4.1.1 120 x 106 Btu/hr Circular Burner
The Circular Burner is B&W's pre-NSPS (New Source Performance Standard)
burner design. The test burner shown in Figure 4-1, was scaled to match the i' 4-
design criteria of the burners at Comanche Unit 2. The Circular burner is a
simple design, with a central coal pipe surrounded by a single concentric
annular secondary air passage. The inlet to the coal nozzle 1s formed by a
90° pipe elbow. An impeller made up of 4 concentric conical rings at a 75°
included angle imparts a radial component to the coal/primary air stream to
enhance mixing between the coal and secondary air streams. The axial
position of the impeller could be varied in relation to the coal nozzle exit
using the impeller support pipe. The Circular burner, as well as all the
other test burners, utilized a steel throat and exit which were water spray
cooled. In actual boiler installations, the burner exit is generally formed
by tube bends in the water wall covered by a thin refractory layer.
The secondary air register was a conventional assembly of adjustable
radial guide vanes. Varying the position of the vanes generated varying
degrees of tangential swirl. A theoretical swirl number can be defined
as5>6:
where Gg, = axial flux of tangential momentum
Gx = axial flux of axial momentum
R = equivalent nozzle radius
The calculation of swirl number using equation 4-1 requires accurate
measurements of velocity and static pressure distributions to be made in a
cross-section of the swirl jet. However, in the absence of these measure-
ments, the swirl number can be predicted directly from register geometry with
reasonable accuracy. For a guided-vane cascade in radial flow such as the
' 4-2-'i
-------
Secondary Air
.Register
Steel Exi t
Impeller
Support Pi pi
x Coal
+ Primary
Ai
Figure 4-1. B&W Circular Burner.
-------
one used in the Circular burner, the swirl
following empirical expression:
number can be calculated using the
A ^
j- <
S = ^	(4-2)
2B
where = ratio of the average tangential and radial velocity components at
the swirl exit (R) and is defined as;
1 tan
.a' = 		: 		—-	(4-3)
1 - f 1 + tan a tan (rr/z}
B = axial width of the register channel
z = number of vanes in a cascade
a = vane angle
zS
f = blockage factor 			(4-4)
2 Rjcos a	¦ -
S = finite thickness, of the vane
Rj = swirler exit radius
From these empirical equations, a swirl number can be related to radial vane
position for the Circular burner as shown in Figure 4-2.
Burner velocity characteristics at nominal full load of 120 x 10® Btu/hr
are summarized in Figure 4-3. The Circular burner typically operates with a
primary stoichiometry (SRp) of 0.30 and an overall stoichiometry (SRt) of
1.15. At these conditions, the primary velocity is 72 f/s and the secondary
velocity is 112 f/s.
4.1.2 60 x IP® Btu/hr Dual Register Burner
The Dual Register Burner is the current commercial low-N0x burner design
offered by B&W. The test burner was scaled down from the burners in opera-
tion at the Wyodak plant. A cross-section, of the ORB is shown in Figure 4-4.
-------
z
s-
s-
o
(LI
•C
I—
0.5
Register Position (Degrees from Open)
Figure 4-2. Swirl characteristics of 120 x 10^ Btu/hr circular
burner register.
4-5
-------
Firing Rate = 120 x 10® Btu/hr.
SRp - 0.30
Secondary.Air Temp = 550°F
Primary Air Temp = 150°F
200
80 -
-M
4-
U
O
a»
	 Coal Pipe Velocity = 72 ft/s
s-
5
20 -
1.0
0.9
1.1
1.2
1.3
1.4
Overall Stoichiometric Air Ratio (SRy)
Figure 4-3. Velocity characteristics of 120 x 10® Btu/hr
Circular Burner.
-------
Outer Secondary
Air
Adjustable
Register
Vanes .
Inner Secondary
Air in
Adjustable
Inner
Damper
Coal
Diffuser
•Adjustable
Spi n
Vanes
I
Inner Secondary
Air
Primary Air + Coal
Outer Secondary
Air
A
Figure 4-4. B&W Dual Register Burner.
-------
The burner design evolved from the concept of using multiple air zones to
allow controlled, delayed mixing of the fuel and combustion air. The DRB
consists of three concentric passages: a central, cylindrical coal nozzle
surrounded by two annular secondary air passages. Coal, transported by
primary air, enters the coal nozzle through a 90° elbow. A bluff body
di ffuser is located at the inlet to the coal nozzle. This diffuser produces
a uniform coal distribution across the coal nozzle without imparting any
swirl or radial component to the primary air/coal stream. The combustion, or
secondary, air is divided between two annular passages. The inner passage is
equipped with an adjustable damper, or sleeve, for flow control and a set of
adjustable axial spin vanes for swirl control. The outer secondary air
passage utilizes adjustable radial register vanes for both flow and swirl
control. During these DMB evaluation tests, the DRB and the DMBs were
Instal led in a compartmented windbox so that the air flow to the inner and
outer secondary air passages could be metered and controlled independent of
burner adjustments.
Swirl characteristics were calculated for both inner and outer zone
devices. The outer zone radial vane assembly could be treated like the
Circular burner register assembly in equation 4-2. For an axial flow swirl
generator such as that used for the inner passage, the following empirical
expression can be used to predict the swirl number
\
tan a \
..J
z = Rh/R
where R^ = hub radius
R = spin vane radius
= spin vane angle
The swirl characteristics of the Dual Register Burner are summarized in
Figure 4-5. Burner velocity characteristies at nominal full load, with an
overall stoic hi ometry of 1.20 and a primary stoichiometry of 0.20, are
4-8
S =
l_z3
1-z2
-------
£ [Inner Spin Vanes ']
H Outer Register
-------
summarized in Figure 4-6. In the general context of swirl characteristics it
should be noted that the DRBs employed in Wyodak Unit 1 are an earlier
version of this burner design. In this version the inner register assemblies
are of the radial inlet design, similar to those employed in the outer air
register. Also, the primary air mixing device is a venturi rather than a
diffuser. This Tatter variable was however shown to have only a small impact
on mixing/NOx emissions in testing conducted on an 80 x 10^ Btu/hr DRB.
4.1.3 60 x 1Q6 Btu/hr Distributed Mixing Burners
B&W integrated the basic DMB design criteria with their commercial
burner components. This adaptation resulted in an arrangement resembling a
Dual Register Burner surrounded by four equally spaced tertiary air ports.
The components common to the DRB include a central, cylindrical coal nozzle
and two concentric, annular secondary air passages. As for the DRB,
pulverized coal with primary air enters the coal nozzle through a 90° elbow.
The inner secondary air passage is equipped with a sliding sleeve damper for
flow control and adjustable axial spin vanes for swirl generation. The outer
secondary passage is equipped with a register assembly of radial vanes for
both flow and swirl control. B&W designed the tertiary air ports with a
divergent, cone-type exit for each of the prototype DMBs.
Three configurations of the half-scale, 60 x 10^ Btu/hr prototype B&W
DMB were evaluated. The two configurations initially evaluated are shown in
Figure 4-7. One of these initial configurations utilized a 75° coal impeller
similar to that used with the B&W Circular burner. This impeller was
expected to produce a "short" flame under staged conditions, thus making
their configuration appropriate for retrofit in boilers with shallow firing
depths. The other configuration incorporated the Dual Register Burner type
coal diffuser. This configuration was expected to produce a longer flame
with. 1 ower N0X emi ssions which coul d be accommodated in larger, post-NSPS
applications. The mounting arrangement on the LWS, common for all three 60 x
3.q6 Btu/hr DMB configurations, is shown in Figure 4-8.
4-ip_.
-------
Nominal Conditions:
£
Firing Rate = 60 x 10 Btu/hr
Air Temp = 5Q0UF
240
200
Inner Secondary
Air^"——.
o
¦
'100
Percent of Secondary Air to Inner Zone
Figure 4-6. 60 x 10& Btu/hr Dual Register Burner velocity
characteristics.
4-11
-------
Tertiary Air Port
Inner Secondary Sleeve
Impeller Support Pipe 	
Steel
Inner
Secondary Sleeve-
Outer Secondary
Coal
Impeller
Register:
Inner Secondary
-Outer
'Secondary
Register
Vanes
Coal ' «
Diffuser;

Inner
Secondary
Vanes
Steel ¦,
ExiJ^l
^ Tertiary Air Portsc
a) 75° Impeller Configuration
b) Coal Diffuser Configuration
Figure 4-7. Initial 60 x 10 Btu/hr DMB configurations.
-------
lO'-O
LWS Mounting Plate
(TYP)
DMB Exit
-©¦
4 Tertiary Air Ports Equally
Spaced on 3'8" Radius
Figure 4-8
Mounting arrangement for 60 x 106 Btu/hr DMBs.
-------
During analysis of data from tests with the initial 60 x 10® Btu/hr 0MB,
outer secondary air velocities were determined to be unusually high.
Investigation of the design parameters revealed that the outer passage was
improperly sized. The velocity characteristics for the initial half-scale
DMB, shown in Figure 4-9, show that under staged conditions at a burner zone
stoichiometry of 0.70, an equal distribution of air between the inner and
outer passages yields an outer secondary velocity of about 240 f/s. This is
significantly higher than the DMB design criteria of 60 f/s for secondary
air.
Following optimization of the properly sized 120 x 10® Btu/hr DMB, the
half-scale DMB was modified to match the design of the full scale, 120 x 10®
Btu/hr DMB. This modified 60 x 10® Btu/hr DMB is shown in Figure 4-10. This
third half-scale DMB configuration utilized a scaled version of the optimum
coal spreader from the 120 x 10® Btu/hr DMB tests. This scaled-down version
of the coal spreader had the following design characteristics;
Support pipe diameter = 2.875 in. O.D.
Blade Height = 3.Q62S in.
Blade Length = 4.4375 in.
Blade Angle = 30° from axial
This design produced an effective swirl number of 0.415, compared to the
full-scale version, described in section 4.1.4, with a swirl number of 0.414.
The velocity characteristics of the modified half-scale DMB are shown in
Figure 4-11.
The same swirl generators were used in the initial and modified 60 x 10®
Btu/hr DMBs, only the outer secondary passage outside diameter {i.e. burner
throat) was changed. The swirl character!"sties were therefore the same.
These are summarized in Figure 4-12. As with the Dual Register Burner, the
radial vane assembly used for the outer secondary passage is a more effective
swirl generation device than the inner passage, axial spin vane assembly.
4-14
-------
Nominal Conditions:
Firing Rate = 60 x 10® Btu/hr
SRt =1.20
SRp = 0.20
Secondary Air Temp = 500°F
500
400
Outer Secondary
Inner Secondary
100
0
20
40
60
80
100
Percent of Secondary Air to Inner Zone
Figure 4-9. Velocity characteristics of initial 60 x 10® Btu/hr DMB
under staged conditions (SRg = 0.70).
"4-15
-------
Tertiary Ai
Port
Outer Secondary
Register
Inner Secondary
v Sleeve
Coal Pipe
Coal Spreader
Inner Secondary
Snin Vanes
e 4-10. Modified 60 x 10 Btu/hr Distributed Mixing Burner.
-------
Nominal Conditions:
Firing Rate * 50 x 10® 8tu/hr
SRt = 1.20
SR = 0.20
p	o
Secondary Air Temp = 500 F
| Ur,staged
i	 Staged	
(SR. • 0.70)
^__b
320
280
240
u
i
- 160
Outer
Secondary
U
o
X
100
Percent of Secondary Air to Inner Zone
Figure 4-11. Velocity characteristics of modified
60 x 10D Btu/hr DMB.
4 -17
-------
14
12
Inner Spin Vanes
Outer Register
10 -
1
I
20	40	60	80
Vane Angle (Degrees from Open)
6
Figure 4-12. Swirl characteristics of 60 x 10 Btu/hr DMB.
4-1-8-
-------
4.1.4 120 x 106 Btu/hr Distributed Mixing Burner
The full-scale, 120 x 10^ Btu/hr Distributed Mixing Burner is shown in
Figure 4-13. The principal components were the same as described previously
for the half-scale DMBs, with DRB-type hardware and four tertiary air ports
with conical divergences. The mounting arrangement onto the LWS test furnace
is shown in Figure 4-14. The operational characteristics for this full-scale
DMB, burner velocities and swirl, are shown in Figures 4-15 and 4-16,
respectively.
With this burner operating in a staged combustion mode it was found
necessary to implement a series of modifications to the coal nozzle to
produce flame dimensions which could be accommodated within the available
firing depth of the LWS. These modifications were developed in an iterative
manner, and consisted of seven variations of coal spreader design. Four
designs, shown in Figure 4-17, were based on the standard B&W 37° impeller
made up of three concentric conical rings around a center cone shaped bluff
body. Variations to this basic design included reducing the number of rings
and lengthening the ring to cover more coal pipe cross-section, Figure
4-17(b); adding six support vanes, set 25° angle between the center body and
conical ring, Figure 4-17(c); and removing the outer ring and using six
larger vanes set at a 25° angle.
Since these designs achieved limited success in reducing flame length,
the conical impeller-type spreader was abandoned in favor of single swirler
designs. The three designs tested are summarized in Figure 4-18. The design
which yielded an acceptable flame length had the following design
characteristics:
Support Pipe Diameter = 4.0625 in.
Blade Height = 4.375 in.
Blade Length = 8.25 in.
Blade Angle = 30° from axial
-------
Tertiary Air
Port
Impeller Support
Pipe
Coal Pipe
Steel Exit
Coal Impel 1
Inner Secondary
Spin Vanes
Figure 4-13. 120 x 10^ Btu/hr Distributed Mixing Burner with standary B&W coal impeller.
-------
10'-0
^ LWS
Mounting
Plate
O
DMB Exit
LO
(TYP)
o
o
4 Tertiary
Ports Equally
Spaced on 510
Radius
Figure 4-14. Mounting arrangement for 120 x 10® Btu/hr DMB.
-------
Nominal Conditions:
Firing Rate « 120 x 106 Btu/hr ,
SRt » 1.20
SR = 0.70
P	0
Secondary Air Temp = 500 F
Unstaged	""	¦¦¦¦
.	Staged	— — — —_ {SRfc = 0.70)
320
280
240
200
U
OJ
W!
Inner
Secondary
160
>.
-------
10
8
~T		
Inner Spin Vanes
/Outer Register
S_
-Q
! £ ''
: *5 '¦
! lO
I to
i <->
! '£ 4
i
I s- ¦
o
-------
Bluff Body
Zone
(a) Baseline 37° Impeller
(b) Dual Cone 37° Impeller
6 Swirl Blades
<51 OCQ	wi
I
(c) Dual Cone 37° Impeller with 6 - 25° Vanes
6 Swirl Blade1
a	av
(d) Single Cone 37° Impeller with 6 - 25° Vanes
Figure 4-17. Impeller configurations tested with 120 x 106
Btu/hr DMB.
4-24
-------
Coal Pipe
4" Suppor
Pipe
Number of
Blade
Blade .
Blades
Angle (X)
Length (L)
4
25
4"
4
30:
6"
4
30
8"
6
Figure 4-18, Swirler designs tested with 120 x 10 Btu/hr DMB.
'4-25
-------
This design produced an effective swirl number of 0.414 and was selected as
the final, optimum configuration of the 120 x 106 Btu/hr DMB.
4.2 Fuels and Sorbents
Four different coals were used for the purpose of this test program:
Utah coal, Illinois coal (high in sulfur), and one coal from each host boiler
site. Data from tests with a fifth coal, Pittsburgh #8, in conjunction with
the LIMB demonstration program, are also included to broaden the
interpretation of results. The compositions of these coals and their
respective ashes are again listed in Table 4-1 and 4-2, respectively. The
Utah coal has been used as the base fuel at EER in the development of low-
emission, high-efficiency burners. It is a high volatile B bituminous coal
from the Western United States with a low sulfur content. The high-sulfur
coal used is Illinois #6 coal. This is a high-volatile C bituminous coal
selected to provide data which would be applicable to eastern U.S. boilers
burning high-sulfur fuels. The Illinois coal has been tested at EER during
previous studies in the LWS, and thus, will permit comparisons with this
program. The Wyodak coal is from the DRB host site, Wyodak Plant. This is a
subbituminous B coal from Wyoming. Testing the DRB with the Wyodak coal will
establish a link to data from an operating boiler. Similar objectives are
gained by testing the coal from the Circular burner host site, Comanche Unit
2. This boiler also uses a Wyoming subbituminous B coal, denoted Comanche
for this project. The fifth coal with data presented in Part I of this
report is Pittsburgh #8. This high-volatile A bituminous coal was the coal
selected for the EPA LIMB demonstration program at Ohio Edison's Edgewater
Unit 4 and was used during the second generation low-NOx burner selection
tests, discussed in Part II of this report.
Predictions of N0X formation and reduction potential based on coal
composition have been developed at EER to rank coal types. Empirical
correlations have been developed in a laboratory combustor for N0X emissions
based on coal properties for several distinct combustion conditions.^
Application of these correlations yielded results summarized in Table 4-3.
The NO predictions listed include theoretical total conversion of fuel
4-26,
-------
TABLE 4-1. COMPOSITION OF TEST COALS
Coal
Utah
Illinois •
Wyodak
Comanche
Pittsburgh #8
Reporti ng
Basis
As
Rec'd
Dry
As
Rec' d
Dry
As
Rec'd
Dry
As
Rec'd
Dry
AS
Rec1 d
Dry
Proximate
{wt. %)










Moisture
Ash
Volatile
Fixed C
6.11
8.02
41,26
44.60
0.00
8.55
43.96
46.73
15.26
8.09
34.60
42.06
0.00
9.54
40.84
49.64
23.85
7.17
33.70
35.29
0.00
9.41
44.22
46.37
22.44
5.00
36.12
37.72
0.00
6.45
44.87
48.68
3.50
12.92
33.75
49.83
0.00
13.40
34.98
51.62
Heati ng
Value










Btu/lb
MMF Btu/lb
mf Btu/lb
12,288
13,088
14,440
14,311
10,710
12,638
14,209
14,088
8,945
11,753
13,085
12,963
9,325
12,026
12,939
12,855
12,177
12,618
14,876
14,626
U1 timate
(*t. %)










Moisture
Carbon
Hydrogen
Nitrogen
Sulfur
Ash
Oxygen*
6.11
68.58
5.16
1.28
0.60
8.02
10.24
0.00
71.86
5.49
1.36
0.64
8.55
10.91
15.26
59.45
4.28
1.07
3.23
8.09
8.64
0.00
70.14
5.05
1.27
3.81
9.54
10.21
23.85
50.93
3.65
0.75
0.43
7.17
13.23
0.00
66.89
4.81
0.98
0.57
9.41
17.34
22.44
54.25
3.80
0.76
0.43
5.00
13.32
0.00
69.97
4.91
0.98
0.56
6.45
17.14
3.50
68.13
4.63
1.21
3.22
12.42
6.41
0.00
70.54
4.79
1.26
3.30
13.40
6.63
Forms of
Sulfur
(wt. %)










Sulfate
Puri ti c
Organic
0.01
0.13
0.46
0.01
0.13
0.50
0.18
0.95
2.11
0.21
1.11
2.49
0.01
0.06
0.36
0.01
0.09
0.47
0.02
0.09
0.32
0.02
0.12
0.42
0.22
1.62
1.38
0.23
1.65
1.42
*Oxygen determined by difference.
4-27'
-------
TABLE 4-2. COAL ASH CHARACTERISTICS
Elemental Ash
Utah
m inols
Wodak
Comanche
Pittsburgh #8
Elemental Ash





(wt. %)





Si 02
58.40
49.03
34.48
23.18
48.67
A1203
19.96
17.71
17.10
13.99
20.19
TiOg
0.77
0.68
0.78
1.04
0.84
^eZ®3
4.18
18.07
5.48
5.07
23.87
CaO
4.56
4.37
19.73
28.42
1.60
MgO
1.05
0.78
5.29
5.15
0.60
Na20
1.54
1.02
1.29
1.20
2.00
K20
1.06
1.91
0.53
0.29
0.31
P205
0.51
0.21
1.03
1.41
0.39
so3
4.77
4.47
12.62
17.50
1.25
Ash Fusion





Temperatures





(°F)





Oxidizing





IDT
2350
2337
2233
2390
2377
ST
2448
2409
2258
2412
2554
HT
2546
2479
2300
2425
2580
FT..
2653
2533
2311
.2451
2616
Reducing





IDT
2297
2041
2163
2316
2171
ST
2388
2135
2203
2342
2298
HT
2502
2310
2214
2351
2459
FT
2621
2339
2272
¦ 2383
2498
4-28
-------
TABLE 4-3. PREDICTIONS OF NO^ EMISSIONS BASED ON GOAL COMPOSITION
Coal
Utah
Illinois
Wyodak
Comanche
Pittsburgh #8
ASTM D388
Rank
High-Volatile
B Bituminous
HI gh-Vol ati le
C Bituminous
Subbi tumi-
nous B
Subbilumi-
nous B
High-Volatile
A Bituminous
Compost -
tion (wt.
% daf)





Nitrogen
1.49
1.40
1.08
1.05
1.45
Volatile
Matter
48.07
45.15
48.81
47.96
. 40.39
Fixed
Carbon
51.10
54.88
51.19
52.04
59.61
NO Pre-
dictions
(ppm @
. 0% 02)





Theoret-
ical
2987
2836
2402
2297
2790
Premixed
1188
1082
935
911
1058
Radial
Diffusion
876
821
708
694
825
Mi ni mum
Staged
. 275
276
244
242
288
4-29_
-------
nitrogen to NO, and NO emissions predicted for different mixing conditions:
a premixed flame, a radial diffusion flame, and physically staged combustion
air conditions for minimum N0X emissions. The absolute values of the NO
predictions are for the specific laboratory combustor and would be expected
to vary with combustor design and operation. High-turbulence, pre-NSPS type
burners would be representative of conditions between premi xed and radial
diffusion. For the subject coals, the Utah coal would be expected to yield
the highest N0X emissions with a conventional burner, such as the Circular
burner while the two subbiluminous coals would produce the lowest N0X levels.
However, the Utah coal is most amenable to staging, with 77 percent reduction
in NQX from premixed to staged conditions. Under staged conditions, such as
those achi eved with a DMB, the two subbi tumi nous coals would again be
expected to produce the lowest N0X and the Pittsburgh #8 coal the highest.
However, the correlation indicates that absolute NGX emission values for the
di fferent coals are much closer under staged combustion conditions, compared
to premixed conditions where differences can be large.
As-fired pulverized coal samples were obtained on a daily basis
throughout the testing period. The pulverized coal was sampled downstream of
the pulverizer exhauster following ASME PTC 4-2 procedures. The objective of
this sampling was to verify the composition and fineness of the coal. The
mean compositions of the different coals, determined from the average values
of all daily samples, are summarized in Table 4-4, which shows also the
relative standard deviation of the individual components of the ultimate
analysis. The Pittsburgh #8 was used briefly during this part of the program
and sampled but once, this standard deviation could not be determined. The
standard deviation for the four main fuels tested was less than 2 percent,
suggesting very consistent coal composition.
Typical particle size distributions for the test coals are shown in
Figure 4-19. Daily variations of coal fineness are shown in Table 4-5. Coal
fineness was maintained nominally at 70 percent through 200 mesh (75Mm),
consistent with industry standards. Actual values ranged from 67 percent for
the single day of testing with Pittsburgh #8 coal to 74.3 percent for the
Wyodak coal.
4-30
-------
T-AfiLEi4-.4._JlEAN COAL COMPOSITION DATA—AVERAGES OF ULTIMATE ANALYSIS
"" PERFORMED ON DAILY COAL* SAMPLES.

Utah
Illinois
Wyodak
Comanche
Pittsburgh #8*
Coal
Mean
Std.
Dev.
Mean
Std.
Dev.
Mean
Std.
Dev.
Mean
Std.
Dev.
Mean
Std.
Dev.
Composi tion
(Dry, wt. %)










Carbon
71.55
1.98
66.81
0.69
64.34
0.77
65.06
1.14
72.00
-
Hydrogen
5.26
0.11
4.67
0.12
4.62
0.11
4.49
0,16
4.73
-
Nitrogen
1.41
0.07
1.26
0.07
1.00
0.03
1.00
0.04
1.39
-
Sulfur
0.63
0.07
3.79
0.08
0.43
0.02
0.55
0.06
3.12
-
Ash
7.86
1.62
9.76
0.46
7.75
0.03
6.52
0.45
10.02
-
*Single day of operation.
-------
99.,9
99.9
99.8
99.5
99
98
95
90
80 >
70
60
50
40
30
20
10 h
!10
x
J	I	L
J	L
X
J	L
X
20
30 40 50 . 6j}„80 10.0.
.200 300 400 500 600 1000
I O	Utah
~	Illinois
A	Wyodak
<0	Comanche
0	Pittsburgh #8
Particle Size {pi rn)
Figure 4-19. Coal particle size distributions.
4-32
-------
TABLE 4-5. VARIATION OF COAL FINENESS-WEIGHT PERCENT
PASSING 200 MESH SCREEN
Pittsburgh #8*
Utah
mi nol s
Wyodak
1 Comanche
Pittsburgh #8*
Mean (wt. %)
70.9
70.6
74.3
72.4
67.0
Std. Dev. (%l
3.96
4.52
4.41
5.93
- p g 1 e.	e s t trig, 4,
4-33
-------
Three sorbents were used during this program to evaluate the potential
of SO2 reduction with in-furnace injection, Vicron 45-3 limestone and Colton
hydrated lime have been used at EER as examples of each type of material in
the development of LIMB technology. In addition, a third sorbent was
evaluated for general interest because of its highly reactive nature-
pressure hydrated dolomitic lime. The physical and chemical characteristics
of these sorbent materials are listed in Table 4-6 and the corresponding size
distributions are shown in Figure 4-20. Vicron 45-3 is nominally 99 percent
pure CaC03 with a mass median diameter of 9.8/Am. The Colton hydrated lime
is nominally 96 percent Ca(OH)2 with a median particle size of 4.0Mm* The
pressure hydrated dolomitic lime is a much finer material, with a mass median
diameter of only 1.4 M in addition to containing a significant amount of
magnesium oxide.
4.3 Test Facility
4.3.1 Large Watertube Simulator
Testing of the burners was conducted in the EPA Large Watertube
Simulator (LWS) at the El Toro, California, test site of EER. The LWS has a
capacity to accommodate up to 150 x 10® Btu/hr input. The furnace is
designed to match the size and geometry of a large industrial or small
utility single-wall fired furnace. Figure 4-21 shows its general
construction and Table 4-7 lists its design parameters and dimensions.
The LWS furnace is 22 ft deep and 16 ft wide. The overall height is
50.5 ft from the hopper to the top. The hopper is located in a concrete-
lined pit so that the test burners are positioned at ground level. The
furnace is arranged in a front-wall fired configuration with the nose
directly above the rear (or target) wall. Single- and four-burner arrays
have been fired in this configuration at over 100 x 10® Btu/hr. It is also
possible to arrange burners for opposed or corner firing.
The shape of the LWS furnace generally matches industrial and utility
boiler specifications. For example, the hopper design and nose angle are
f 4-34;
-------
TABLE 4-6. PHYSICAL AND CHEMICAL PROPERTIES OF SORBENTS
L- - - . __ 	;	. /
Physical Properties
Elemental Ash %)
Sorbent
Theoretical
Characteri sties
Median
Di aneter
( m)
Densi tv)
(gm/cnw)
LOI 0
1000°C
(wt.%)
CaO
Fe203
A12O3
Na20
MgO
K2O
Si O2
Ti02
P2O5
SO3
Vicron 45-3
Hydrated Lime
Pressure Hydrated
Dolomitic Lime
CaC03
Ca(0H)2
Ca(0H)2-Mg(0H)2
9.8
4.0
1.4
2.706
2.279
2.289
42.49
22.91
N.A.
55.64
72.67
42.42
0.08
0.15
0.10
0.03
0.40
0.49
0.01
0.01
N.A.
0.54
0.42
26.4
0.01
0.06
N.A.
0.20
7.06
0.28
0.01
0.02
N.A.
0.01
0.01
N.A.
0.02
0.07
N.A.
N.A. - No Analyses
Available
-------
00
so
80
;70
60
50
SO
3Q
20
10
0
Col ton
Hydrated Lime
Vicron 45-3
Limestone
Pressure Hydrated
Dolomitic Lime
J	j_

_L
_L
J	L
J	L
1
_L
)0 60 50 40 30 20	10 8 6 5 4 3 2
Equivalent Spherical Diameter (^m)
Figure 4-20. Particle size distributions of sorbent materials.
1 0.8 0.6 0.4 0.3
-------
To
Scrubber
Observation
Sample Port
(Typ)
Slagging Test
Panel (Typ)
Access Door
(Typ)
Safety
Vent
Purlins
Q 0
Burner Mounting Plate
Figure 4-21. General configuration of Large Watertube Simulator,
LWS.
4-37
-------
TABLE 4-7. LWS FURNACE CHARACTERISTICS
Parameter
Value
Firing

Configuration Primary
Single Wall
Optional
Opposed or Corner
Firing Rate Minimum
50 x 106 Btu/hr
(Practical Limit)
Maximum
150 x 106 Btu/hr
Dimensions

Firing Depth
22 ft
Width
16 ft
Rear Wal 1 Hei ght
20 ft
Knuckle-to-Nose Height
27.5 ft
Nose Angle
37° (From Horizontal)
Hopper Angle
52° (From Horizontal)
Height Above Ground Level
35.7 ft
Total Vertical Height
50.5 ft (Pit-to-Top)
Cooled Surface Area

Burner Zone
804 ft2
Total Furnace
1951 ft2
Volume

Burner Zone
3520 ft3
Total Furnace
11,544 ft3 (to Nose)
Insulation
All Sidewal 1 s Knuckle-
Up 16 ft
Cooling
Spray-Cooled on All
Surfaces
4-38
-------
Identical to those used for field operating equipment. The area above the
nose would normally contain superheater tubes in a field installation. In
the LWS, this area is empty. However, the area has been sized so that if the
convective tubes are installed, the tube size, spacing, and gas velocity will
approximate typical field-operating boiler specifications.
The LWS furnace is externally spray-cooled with water to absorb the heat
of combustion and control furnace wall temperatures. The vertical walls of
the furnace in the vicinity of the burner are insulated with refractory so
that furnace internal temperatures are similar to those of field operating
boilers. The four sidewal1s are insulated with 2 inches of Kaiser I-R-C
refractory from the hopper knuckle up to 16 ft. The front and back slope of
the hopper are insulated across the width of the furnace and 8 feet down the
hopper si opes with one 2.5-i nch layer of G-26 firebrick. The heat 1ibera-
tion/cooled surface area is also similar to field-operating boilers at
nominal full load.
An overall view of the LWS system is presented schematically in Figure
4-22. Combustion air supplied by a forced-draft blower passes through a
tubular heat exchanger and into a manifold. The combustion air flow rate is
measured and controlled by multiple Venturis and dampers. For burners
equipped with a common windbox, several Venturis can be connected in
parallel. The use of multiple Venturis allows accurate flow rate measurement
over a wide flow rate range. The other side of the heat exchanger is
supplied with hot exhaust from a separate oil-fired combustion chamber. This
allows the combustion air preheat temperature to be controlled independently
of the performance of the burner firing in the research furnace.
The fuel supply system can handle liquid and solid fuels.	Underground
tanks and pumps can be used to fire a wide range of liquid fuels.	Compressed
air and steam are available for atomization. Two Raymond bowl	pulverizers
are available for firing pulverized coal directly. Alternatively, coal may
be pulverized and removed from the primary coal/air stream in	a baghouse.
The pulverized coal is stored in hopper bottom bins and can be	utilized to
form coal/oil or coal/water mixtures if desired.
I-39-
-------
1.
Coal Storage
7.
z.
Main Coal Hopper
8.
3.
Crusher
9.
4.
Coal Feeder
LQ.
5.
Pulverizer
11.
6.
P.A. Exhauster
12.
Main Combustion Air Fan
Tubular Air Heater
P.A. Trim Heater
P.A. Venturi
Oil/Gas Fired Preheater
Air Plenum
13.	Combustion Air Venturis
14.	Burner
15.!Windbox
16.	LWS
17.	Exhaust Duct
18.	Spray Tower Scrubber
Figure .4-22; Large Watertube Boiler Simulator facility.
-------
The furnace exhaust system was designed to permit a wide range of
emission measurements. The sampling location is near the end of a long
straight duct meeting EPA specifications for the minimum number of sampling
points. Since there is no convective section, the temperature at the
sampling point is about 1200°F. The exhaust duct is externally spray-cooled
like the furnace. Downstream of the sampling location, the exhaust passes
into a calcium carbonate scrubber which controls S0X and particulate
emissions. This system meets all applicable air pollution control
regulations.
4.3.2 Test Configuration
One of the key aspects of this restructured program was to evaluate the
effects of furnace environment on burner performance and to compare these
effects with those of a full-scale utility boiler. Since the LWS geometry is
fixed, the criteria used to modify the furnace environment was to match the
mean furnace exit temperature at 120 x 1Q6 Btu/hr while firing the reduced
scale 60 x 10® Btu/hr burners through the use of additional insulation. The
insulation requirements were evaluated by EER using the Richter furnace heat
transfer model.
Several iterative trials were run for selected insulating configura-
tions. Additional insulation in the distribution shown in Figure 4-23 was
considered. Results from the modeling work indicated that by covering the
area with a ceramic-board type of insulation, exit temperatures for 60 x 10®
Btu/hr input could match the exit temperatures with the baseline insulation
configuration fired at 120 x 106 Btu/hr.
The success of this insulation and thermal furnace exit temperature
match is summarized in Table 4-8. Furnace exit gas temperatures measured
with suction pyrometers, or high velocity thermocouples shows the
effectiveness of the insulation pattern. At 120 x 10® Btu/hr, the average
furnace exit temperature was 1810°F with no insulation. Firing at 60 x 10®
Btu/hr, the exit gas temperature in the uninsulated furnace was 250°F lower,
4-41
-------
r -p-
i ¦ i
I i -E*
1 ro
Add7 tional :¦
>: Insulation:
•: (SjJeckled) :
Existing
-ZZ-Z Refratiory Z-_-
o o
Uni nsulated
Wountinn
Opiate O
mmm
Existing
F1 rebri ckSsi
16'
f
Additional
Insulation
(Speckled)
Existing
Refractory
Refractory
Existing
Fi rebri ck
;Addi tional I nsul a -
j.on (Speckited)';£
Existing
	Refractory

ftWfSS
Existing
Fi rebri ckgiiii
11
SSSffiK
Rear Wall
¦SSsSSiSS Additional ;i
I n S u 1 a t i 0 n::
ssbs: (Speckled) 5
Existing "
Refractory
Refractory
North Side Wall
Front Wall
South Side Wal1
Figure 4-23. Insulation pattern in the LWS for operation at 60 x 10 Btu/hr.
-------
TABLE 4-8. LWS FURNACE EXIT GAS TEMPERATURES

BURNER
FURNACE
EXIT TEMPERATURE (°F)

CONFIGURATION
AVERAGE
RANGE
120 x 106 Btu/hr DMB
Uninsulated
1792
1723-1878
120 x 10^ Btu/hr Circular
Uninsulated
1828
1770-1912
60 x 106 Btu/hr DMB
(Initial Design)
Uni nsulated
1562
1484-1640
60 x 106 Btu/hr DMB
(Initial Design)
Insulated
1776
1732-1819
60 xJO6 Btu/hr DMB
(Modified)(
Insulated
1776
1657-1828
-------
with the insulation pattern shown in Figure 4-23, the average furnace exit
gas temperature was 1776°F, within 35°F or 2 percent of a perfect match.
4.4 Test Procedures
Established test procedures were utilized during the evaluation of the
DMB concept in the LWS. The test procedures were in accordance with
guidelines set by previous EPA Quality Assurance Project Plans and described
in the Burner Evaluation Test Plan submitted June 1984. Specific quality
assurance activities for these tests are documented in Part V, Appendix A of
this report. A routine set of input/output measurements was completed for
all test conditions. These measurements document all input parameters which
specified the test conditions and monitored the key overall performance
parameters including flame stability and emissions. These standard
measurements are listed in Table 4-9. Most of the measurements were made
continuously and results were processed into engineering units in real-time
by a microcomputer. Those parameters which must be recorded manually were
entered into the computer separately so that the computer generated data
record is a complete listing of all parameters. These computer listed raw
data are included in Part V, Appendix B.
The sampling train utilized for the continuous emission monitoring of ^ ,s
the flue gas is shown in Figure 4-24. All materials in contact with the
sample are glass, 316 stainless steel or Teflon. These materials are
nonreactive with NO at low temperatures. The stainless steel sampling probes
are water cooled, and small cyclones adjacent to the probe remove particulate
matter and condensed water. The sample is then transported to the control
room through a Teflon line where it is filtered and dried. The sample pump
is Teflon lined and the sample flow rate is maintained at several times the
instrument's requirements to minimize response time. Excess sample flow is
bypassed to a vent. Commercial gas analyzers are used for measurements of
O2, CO2, CO, N0/N0x, and SO2 and are listed in Table 4-10. The accuracy of -f p.
each instrument is maintained by frequent calibration with zero and certified
calibration gases.
4-44
-------
TABLE 4-9. STANDARD INPUT/OUTPUT MEASUREMENTS/
Parameter
Measurement Method
Record
Frequency
Burner Settings
Observation
Manual
Each Test
Fuel Flow Rate
Weigh-Belt Feeder and
by Oxygen Balance
Calculations
Computer
Continuous
Fuel Fineness
ASME PTC 4.2
Manual
Once/Day as
Requi red
Fuel Composition
Samples Obtained
Prox. & Ult.
Analysis
Manual
Manual
Once/Day
Selected Tests
Combustion Air
Flow Rate
Calibrated Venturi
Computer
Continuous
Combustion Air
Temperature
Thermocouples
Computer
Continuous
Flame Charac-
ter i sties
Direct Inspection. ¦
Color Video
Manual
Monitor/Tape
Once/Test
Continuous
Exhaust
Composition
Analyzers
02—Paramagnetic
CO—NDIR
CO2—NDIR
N0x--Chemi1umi nescence
Strip Chart
and Computer
Conti nuous
Furnace Exit
Temperature |
Suction Pyrometer
Manual
Selected
Tests >
4-45
-------
Sample Connections
for LWS and'SMS
Exhaust
Duct
Water-Cooled
Stainless
Steel Prob
Stainless
Steel
Cyclone
Drain Valve
A
Bypass C
Rotameter
Bypass Con-
trol Val
refiIter
Pressu
Gauge
Secondary
Gauge Filter
Teflon-Lined
Diaphragm ^Pun^jj^
Stainless Steel
Refrigerated
Water Trap
Selection
Valves
Teflon
Sample Line
to Blockhouse
Rotameters
?yPass Metering
S^ple Va1ves
^
^ ^
^ ^

©
«>
-<8>
n
ja
cr
a
to
CO
<1>
M
c
(O
a.
tn
5
o
c
tO
Q_
CO
iIters
tn>
Gas Analysis
Instruments
O2
C02
CO
N0/N0v
S02
Figure 4-24. Flue gas sample train schematic.
-------
TABLE 4-10. GAS PHASE SPECIES INSTRUMENTATION
Species
Operating Principle
Instrument
Model No.
02
Paramagnetic
Beckman
755
CO 2
Nondi spersi ve
Infrared
Anarad
AR-6QQ
CO
Nondi spersi ve
Infrared
Anarad
AR-5Q0R
NQ/NQX
Chemi1umi nescence
Teco
10R
so2
UV Absorption
DuPont
400
4-.47J
-------
For selected test conditions, fly ash samples to determine the extent of
unburned carbon were collected.. Figure 4-25 shows the system used for this
purpose. The system consists of a 3-inch 1.0. nozzle, insulated stainless
steel cyclone, orifice flowmeter, and an induced draft fan. A large 8-inch x
10-inch glass fiber filter holder also can be connected in series or in
parallel with the cyclone. For normal conditions, the cyclone collects the
fly ash at a rate of approximately 50 grams in 5 minutes sampling time. The
cyclone itself is designed with a cutoff diameter (D50) of about 3 microns.
For selected conditions, additional measurements were made to more fully
characterize burner operating conditions and burner performance. These
additional measurements are listed in Table 4-11.	1 ¦
4-48
-------
Sample
Nozzle
Exhaust Flow
A1r
lock
Compressed Air
Quick Connects
S ~ T
Filter Holder
AP Manometer
I.D. Fan
Cyclone
Control
Valve
— O
Vent
Orifice -
Meter
Cyclone & Filter {in series)
Figure 4-25. High volume isokinetic particulate sampling system.
4-49
-------
TABLE 4-11. DETAILED MEASUREMENT FORMAT
Inputs
Outputs
Parameter
Method
Parameter
Method
Coal Composition
Ultimate and Proximate
Particulate Matter:


Analysis—ASTH
Total Mass
EPA Method 5


Composition
Ultimate Anal-



ysi S--ASTM
Coal Size Dis-
ASTM Method D197
Combustion Effi-

tri buti on

ciency:

Sorbent Compo-

CO, C02
Standard Mea-
si ti on
ASTM Methods
surement



Format
Sorbent Size
SediGraph Particle
Particulate
Ultimate Anal-
Distribution
Size Analyzer

ysis


Calcium



Utilization:



S02
Contlnuous


Monitor UV



and/or EPA



Method 6


S03
Control led


Condensa-



tion and EPA



Method 8


Particulate
Ultimate Anal-



ysis


Slagging Char-
Water-Cooled


acteristics
Slagging



Panels


Fouling Char-
Fouling Probes


acteristics

4-50
-------
5.0
BURNER PERFORMANCE AND N0X EMISSIONS
5.1	Circular Burner
The 120 x 1Q6 Circular Burner was initially tested using Utah coal. The
baseline settings were based on the information supplied by B&W for design
point operation. The resulting flame was very.long {>22 feet) and wide
(>16 feet) with correspondingly low N0X emissions, which was uncharacteristic
of the expected Circular Burner performance. Parametric adjustments were
made to the available burner controls, including; decreasing the primary air
velocity, retraction of the impeller, and adjustments to the secondary air
register. These adjustments did not produce any significant effects on the
flame shape or emissions. In addition baffles were installed in the windbox
to create an air flow distribution in the windbox similar to field
installation. This also had negligible effect on burner performance. Since
these attempts to alter burner performance had little or no effect, the
approximate design point operating conditions as shown below, were utilized
for all the remaining tests. These tests were conducted after verification
that the excessive flame length would not damage the structure of the LWS.
AIR VELOCITIES (ft/sec)	BURNER SETTINGS
Primary	Secondary	Impeller ^Registe/^
70	110	-1 inch 15° Open
[from zero
position)
The effects of excess air on the Circular Burner N0X and CO emission
characteristics are summarized in Figure 5-1 for both full and 55 percent
firing capacities. At full load, N0X emissions decreased 6 ppm/percent
excess air over the range tested. The flame was stable with no detachment,
with lengths > 22 feet and widths > 16 feet for all conditions. At the
design point overall stoichiometry (SRj) of 1.15, N0X emissions were about
5-1
-------
j
600
500 "
400 -
1 t
| "O
! ^300
i °
! **
o
100 -

Firing Rate
(x 10 Btu/hr)
o
121.62 + 0.82
o
66.82 ! 2.01
1.0 1.10 . 1.20 1.30 1.40
	 Overall Stoichiometry - SRy
120
\
o 40
1.0 1.10 1.20 1.30 1.40
Overall Stoichiometry - SRj
Figure 5-1. Summary of 120 x 10 Btu/hr Circular Burner with Utah coal.
-------
350 ppm.* CO was generally stable throughout the range tested at 40 ppm.
Excess air had similar effects at reduced load {67 x 10® Btu/hr), with a
decrease in N0X of 6 ppm/percent excess air. The flame was slightly less
stable at this reduced firing rate and shorter (18-22 feet). At SRj = 1.15,
N0X emissions were 175 ppm resulting in a 50 percent reduction from full load
operation. CO emissions were about 40 ppm down to an excess air level of 19
percent increasing to 57 ppm at 13 percent excess air.
The effect of excess air on the Circular burner firing Illinois coal is
summarized in Figure 5-2 for two firing capacities. At normal full load
(120 x 106 Btu/hr) N0X emissions decreased by 4 ppm/percent excess air, with
a baseline of 370 ppm. The flame was similar to that of Utah, with lengths
> 22 feet and widths > 16 feet. CO emissions were stable at 35 ppm over the
range tested. Reduction of firing rate to 67 x 10e Btu/hr resulted in
approximately a 36 percent decrease in N0X emissions at baseline conditions
to 235 ppm with a slight increase in CO to about 42 ppm.
A series of tests were conducted firing the Comanche Coal that is used
at the Comanche Unit Generating Station, which is equipped with BAW circular
burners, in order to provide a direct comparison between the burner
performance on the LWS test facility and a field installation. Figure 5-3
summarizes the effect of excess air at two firing capacities with Comanche
coal. At full capacity, N0X emissions decreased 11 ppm/percent excess air
with a design point (SRj = 1.15) of 375 ppm. CO was stable at about 35 ppm
down to an excess air level of 10 percent with a rapid increase beyond that
level . The flame was more characteristic of the Circular burners operating
in the field with lengths of 16-20 feet and widths of 10-12 feet over the
range tested. At reduced load of 70 x 10® Btu/hr excess air. Reduction in
firing rate by 42 percent resulted in a decrease in N0X emissions of
approximately 30 percent down to 260 ppm at SRj - 1-15 with no significant
change in CO emissions.
*Unless otherwise stated, emission values referenced in the text are reported
as corrected to OS 02, dry.
5-3
-------
Nominal Conditions:
Fuel - Illinois

Firing rate
(x 10® Btu/hr)
o
120.7 t 1.51
A
66.62 + 2.06
1.0 1.10 1.20 1.30 1.40
Overall Stoichiometry - SRy
\ .10 1.20 1.30
Overall Stoichiometry - SRT
1.40
LJ
Figure 5-2. Summary of Circular Jurner performance with
111inois coal.
-------
NOMINAL CONDITIONS:
FUEL - COMANCHE
!

FIRING RATE
(xia6 Btu/hr)
i
o
121.19 * 2.58
i
j
A
70.86 ! 1.60
or
i
cn
600
500 -
400 -
t-
XJ
X
o

/
120
.159 ppm
300 _
200 -
100 -
1.00 1.10 1.20 1.30 1.40
Overall Stoichiometry - SRy
1.00 1.10 1.20 1.30
Overall Stoichiometry - SRj
1.40
1
4-»
5
C
o
L-
' to
o
c
>>
¦— • ...
Figure 5-3,. Summary of 'Gircular^Burner performance with Comanche coal.
-------
Thus, while Circular Burner performance with Utah and Illinois coals is
comparable, with the subbituminous Comanche coal the visible flame was much
more compact, and N0X emissions were much more sensitive to increasing excess
air. This behavior is believed to be due to the higher inherent reactivity
of the subbituminous coal, which leads to earlier and more intense heat
release. Consequently, under nonstaged conditions, this leads to better fuel -
nitrogen/oxygen contacting in the early stages of combustion, compared to the
less reactive bituminous coals.
At the Comanche Generating Station, normal operation includes a
percentage of the combustion air going through N0X ports (overfire air ports)
located above the burners. Therefore, to quantify the effect of this
overfire air on burner performance, a brief test was conducted using overfire
air. A percentage of the total combustion air was diverted through four
overfire air ports located 19 feet above the burner centerline. Figure 5-4
summarizes the results. N0X emissions decreased by 7 ppm/percent
stoichiometric air diverted through the overfire air ports from about 450 ppm
to 346 ppm with no change in CO emissions. Flame length increased from about
18 feet at baseline to 22 feet at the maximum overfire air level (15
percent).
5.2	Dual Register Burner
A major objective of the tests with the 60 x 10® Btu/hr Dual Register
burner was to obtain performance data in the LWS for comparison to field data
from a DRB-equipped boiler, the Wyodak Plant in Wyoming. This data would be
used to extrapolate LWS burner performance to practical applications. For
these tests the LWS was insulated to provide a thermal environment comparable
to DMB and Circular burner tests at 120 x 10® Btu/hr. The tests of the
60 x 106 Btu/hr DRB included a brief series of burner adjustments to verify
normal operating ranges while firing Utah coal at full load, followed by
characterization of the DRB over load and excess air variation with Utah,
Wyodak, and Pittsburgh #8 coals.
-------
Nominal Conditions
Fuel - Comanche	6
Firing Rate = 120.6 + 1.18 x 10 Btu/hr
SRt = 1.20
<_n
i
600
Fn0 -
: 400 _
300 "
200 -
100 _
1.0 1.05 1.10 1.15 1-20
Burner Zone Stoichometry (SRg)
1.0
1.05 1.10 1.15 1.20
Burner Zone Stoichometry (SRg)
Figure 5-;;4'. Effect of OFA on .Circular Burner performance.
-------
The main burner adjustments available for the DRB were inner spin vane
position and outer register position. At full load with Utah coal and an
overall stoichiometry of 1.20, N0X emissions ranged from 253 to 392 ppm (on a
dry basis, corrected to 0 percent O2) with corresponding flame lengths of
over 22 feet to about 16 feet. For this burner, N0X emissions were found to
be most sensitive to the setting of the inner spin vane position. Varying
this setting from 40°CCW to 40°CW resulted in N0X emissions ranging from 253
to 392 ppm. For a fixed inner vane setting of 30°CW, changing the outer
register setting from 50°CW to 2Q°CW was found to reduce the N0X emission
from 392 to only 370 ppm. This behavior is different from that observed with
the 78 x 10® Btu/hr DRB tested in Phase V of the program, where outer
register adjustment was found to be dominant. The reasons for this
difference in behavior are not apparent from aspects of burner design or
operation.
The burner settings chosen for characterization with the different coals
were:
•	Inner spin vane = 30° CW
•	Outer register = 40° CW
These settings were selected since they are representative of the burner
settings employed at the Wyodak Plant,
The effect of excess air on DRB performance at full load for the three
subject coals is summarized in Figure 5-5. Lowest N0X emissions were
measured for the subbi tumi nous Wyodak coal and the highest for the high
volatile bituminous Pittsburgh #8 coal. N0X emissions at an overall
stoichiometry of 1.20 were 290, 390 and 415 ppm for Wyodak, Utah, and
Pittsburgh #8 coals, respectively.
For the bituminous coals, unburned carbon in-the fly ash was considered
to be acceptable at nominal operating conditions, but was observed to
increase strongly as excess air was reduced, as shown in Figure 5-5. For the
Wyodak coal, however, the carbon in ash was exceptionally low at nominal
5-8
-------
Nominal Conditions: fi
Firing Rate = 60 x 10 Btu/hr
O Utah
Q Pittsburgh #8
A Wyodak
8.0
A
ffs
\
1.00 1.10 1.20 1.30 1.40 1.50
Figure 5-5. Effect of excess air on 60 x 10^ Btu/hr Dual Register {Burner performance4
at full load.
-------
excess air levels, and generally commensurate with values measured at the
Wyodak boiler {approximately 0.2 percent).
The performance of the 60 x lO6 Btu/hr DRB at 75 percent load is
summarized in Figure 5-6 for Utah and Wyodak coals. At this load, N0X
emission's were virtually identical for the two coals. N0X emissions were
about 200 ppm at an overall stoichiometry of 1.20 for both coals. Again, the
Wyodak coal yielded low levels of carbon in the fly ash even down to an
overall stoichiometry of 1.01. Unburned carbon in the Utah coal ash was
higher at this reduced load condition than that for full load operation.
Additional Dual Register Burner test data in the LWS was collected
during tests of 78 x 106 Btu/hr second generation low N0X burners, as
described in Volume II of the report on this project. Data representative of
typical DRB performance was collected during B&W sponsored testing (B&W P.O.
635-0A0Q8408DM) of a 78 x 10® Btu/hr Phase V DRB. These tests were conducted
in the LWS with additional insulation, yielding an average flue gas exit
temperature of 1855°F. Results from five configurations of the Phase V DRB
tested with Pittsburgh #8 coal are shown in Figure 5-7. Two of the
configurations were representative of commercial applications, the coal
diffuser and the coal pipe venturi. With burner settings representative of
field use, as in the Wyodak boiler, the diffuser configuration produced
293 ppm N0X with a flame over 22 feet long while the venturi produced 350 ppm
N0X with a 20-21 feet flame. At similar flame lengths, the N0X difference
was less pronounced, showing approximately 12 percent higher N0X for the
venturi configurations opposed to the diffuser.
5.3	120 x lQ^ Btu/hr Distributed Mixing Burner
The primary focus of initial screening tests conducted with this burner
was to develop an impeller and/or burner operating parameters that resulted
in a flame that was capable of being staged without severe flame impingement
on the furnace rear wall. The initial burner settings were determined from
the DMB 60 x 10® Btu/hr burner tests with the inner register and outer
registers set at 30° and 10° open, respectively, which produced counter
5-10
-------
Nominal Conditions:	Fuel-
Firing Rate = 45 x 10^ Btu/hr	Q Utah
A Wyodak
I
8.0
800
700
7.0
600
6.0
K 500
5.0
4.0
400
i o
i
I o
^ 300
3.0
! O
I
l
200
2.0
I
100
1.0
0
1.50
1.20
1.00
1.00
1.40
i
1.30
1.20
1.10
srt	srt
Figure 5-6. Effect of excess air on 60 x 10^ Btu/hr Dual Register burner performance
at 75 percent of full; load.
-------
Nominal Conditions:
Fuel: Pittsburgh #8
Firing Rate: 78 x 10^ Btu/hr
Configuration:
Burner Settings: O w/Diffuser
r0 .. D w/Venturi
Inner = ?6 CW
rO,
A w/ASP/Oiffusion
Outer = 35"CW W/FSR/Diffuser
Sleeve = 100% w/asp/FSR Diffuser
Open	

SRi
1	"•
Figure 5-7. Effect of excess air on Phase V DRB-configurations' ^performance.
-------
current flows. The outer register generally had the greatest effect on flame
length with a tightly closed setting resulting in a decreased flame length.
All the impellers were tested at the B&W zero or baseline position and
adjustments were made in both directions to determine the effects on flame
length. The following iterative modifications were made to the impeller to
obtain a reasonable flame length.
1.	B&W Baseline Impeller (@37°).
2.	Dual cone design with no swirl blades.
3.	Dual cone design with 6 blades @ 25°.
4.	Large single cone with 6 blades @25°.
5.	4" support pipe with four 4" blades @25°.
6.	4" support pipe with four 6" blades @ 30°.
7.	4" support pipe with four 8" blades § 30° with an effective twist.
The final design resulted in unstaged flames of about 16 feet in length.
At staged conditions (SRg = 0.70) the flame length increased to approximately
22 feet. These resultant lengths were acceptable for continued operation.
The final optimization test involved changing the direction of spin on
the inner register. Due to high windbox pressures of about 10 and 5 inches
of HgO respectively on the inner and outer secondary passages, it was not
possible to make register adjustments while running the burner. Therefore,
since flame length was the major concern, two alternate positions were tried
to determine the optimum performance.
The optimum configuration of the 120 x 106 Btu/hr 0MB was determined to
be:
• Spreader Design = 4 inch support pipe with four 8 inch blades at a
30° angle from axial.
§ Burner Settings: Inner Spin Vanes = 35° Open clockwise
5-13
-------
Outer Register = 10° Open Clockwise
Spreader Position = 3 in. Retracted
These settings were utilized for all the 120 x 106 Btu/hr DMB performance
characterization tests.
A series of tests was conducted with the two primary fuels, Utah coal
and Illinois coal to determine the effects of excess air at unstaged
operating conditions. The results are summarized in Figure 5-8. The effect
of excess air is similar for both coals, with the Utah coal resulting in a
reduction in N0X of 12 ppm/percent excess air compared to 10 ppm/percent
excess air for Illinois coal. Design point conditions at an overall
stoichiometry (SRj) = 1.20 yielded 640 ppm and about 625 ppm N0X for Utah arid
Illinois coals, respectively. Illinois coal produced slightly higher CO
emissions, approximately 43 ppm compared to about 36 ppm with Utah coal.
Figure 5-9 indicates the effect of staging on the DMB emissions. N0X is
reduced by approximately 58 percent at SRb = 0.61. However, staging also
increased the flame length from about 16-17 feet at unstaged conditions, to
over 22 feet at SRb = 0.61. Since this flame length is unacceptable for
continued operation in the LWS, staging was decreased to SRg = 0.70. This
condition resulted in a 53 percent N0X reduction with acceptable flame
lengths of 21-22 feet. CO emissions were generally unaffected by staging.
The baseline values at optimum conditions (SRb = 0.70, SRj = 1.20) were
300 ppm and 44 ppm for N0X and CO, respectively. The effect of excess air at
staged conditions was also evaluated, and results are shown in Figure 5-10.
N0X emissions were reduced by about 4 ppm/percent excess air with no
significant change in CO emissions. However, carbon in ash was measured to
increase from 2.8 percent unstaged to approximately 5.5 percent at SRb = 0.7.
The DMB performance was also evaluated with two host coals, Comanche and
Wyodak, and with Illinois coal. Figure 5-11 summarizes the effect of staging
with these coals. The greatest effect was with the Wyodak coal, with a
55 percent decrease in N0X as a result of staging to SRg = 0.71, with no
significant increase in CO. At SRb = 0.71, flame lengths were 22 feet,
5-14
-------
30C -
120
100 -
E
Q,
o_
>»
S-
-a
O
o
o
NOMINAL CONDITIONS:
FUELS
O UTAH - Firing Rate = 120.5 1 1.7x10® Btu.hr
~ ILLINOIS - SRj = 1.19 + 0.06
1.0 1.10 1.20 1.30 1.40
OVERALL STOICHIOMETRY - SR-r
1.0 1.10 1.20 1.30 1.40
OVERALL STOICHIOMETRY - SRT
5-,
Figure 5-8. Summary of 120 x 10 Btu/hr DMB unstaged operation.
-------
NOMINAL CONDITIONS
FUEL - UTAH	fi
FIRING RATE = 122.0 1 1.3 x 10 Btu/hr
SRT = 1.20 + 0.01
700
cn
: •
ov
F 400
"O
s-e 300
o
>» 80
TV-hod
0.60 0.70 0.80 0.90 1.00 1.10
BURNER ZONE ST0ICHI0METRY - SRg
0.60 0.70 0.80 0.90 1.00 1.1(
BURNER ZONt blOICHIOMETRY - SRB
Figure 5-9. Effect of staging on 120 x 10 Btu/hr DMB with Utah coal." J
-------
t
NOMINAL CONDITIONS;
en
-J
FUEL - UTAH
FIRING RATE = 121.1 + 1.4 x 10 Btu/hr
SRg = 0.70 + 0.015
S00
CL
CL
¦o
o
X
o
l.o 1.10 1.20 1.30 1.40
OVERALL ST0ICHI0METRY - SRj
120
100 -
80 -
60 -
CL
CL
T3
C\l
o
o 40
O
<_)
20 -
1-0 1.10 1.20 1.30 1.40
OVERALL ST0ICHI0METRY - SRT
6~,
1
Figure 5-10. Effect of excess air on 120 x 10 Btu/hr DMB with Utah coal.
-------

FUEL
A
COMANCHE
O
WYODAK
~
ILLINOIS
FIRING RATE = 120.2 t 1.6 x 106 Btu/hr
SRt = 1.20 t 0.01
0.60 0.70 0.80 0.90 1.00 1.10
BURNER ZONE STOICHIOMETRY - SRB"
140
~ nC
0.60 0.70 0.80 0.90 1.00 1.10
BURNER ZONE STOICHIOMETRY - SR0
&
Figure 5-11.	Effect of staging on 120 x 10 Btu/hr DMB with alternate host fuels.
-------
therefore, no further staging was attempted. Overall N0X emissions were
lowest using the Wyodak coal under optimum conditions, with N0X = 270 ppm and
about 48 ppm CO.
With Comanche coal, N0X emissions were decreased by 53 percent as a
result of staging to SRg = 0.65 with only a slight increase in CO emissions.
Below SRg = 0.65, the flame increased to over 22 feet and N0X was reduced by
approximately 3 percent. At design point SR|j = 0.70 and SR| = 1.20,
emissions were 300 ppm N0X and 45 ppm CO. N0X emissions were similar to Utah
coal emissions, and about 30 ppm higher than achieved with the Wyodak coal.
With Comanche coal the carbon in ash remained at low level s throughout the
staging range.
Staging had the least effect on Illinois coal resulting in a 44 percent
reduction to SRg = 0.71 with no increase in CO emissions. At optimum
conditions, N0X emissions were highest with Illinois coal at 340 ppm. This
is 40 ppm higher than Utah and Comanche coals and about 70 ppm higher than
Wyodak fuel.
Figure 5-12 summarizes the effect of excess air for operation on the
alternate fuels under staged conditions. The greatest effect was with
Illinois coal with a 4 ppm/percent excess air reduction in N0X with a slight
increase (less than 10 ppm) in CO. The effect of excess air with Comanche
and Wyodak coals were similar but less pronounced than Illinois or Utah
coals, with about a 3 ppm/percent excess air reduction in N0X and no increase
in CO emissions.
Further tests were conducted to determine the turndown capabilities of
the 120 x 106 Btu/hr DM8, and Figure 5-13 shows the effect of load reduction
on unstaged operation with Utah and Illinois coals. Results are similar,
with a reduction in N0X of about 21 percent for both fuels with no increase
in CO emissions. With Utah coal, the minimum firing rate was 72 x 106 Btu/
hr, which represents a turndown capability of about 40 percent. The flame
was stable and decreased in length to about 14 feet with N0X emissions of
454 ppm. With Illinois coal, the minimum firing rate was 65 x 10® Btu/hr or
t 5-19
-------
NOMINAL CONDITIONS:

FUEL
A
COMANCHE
O
WY0DAK
~
ILLINOIS
600
en
i
r\3
jo ;
g; 400
s-
¦Q
CM
° 300
5-e
o
X
o
FIRING RATE = 120.6 + 1.46 x 106 Btu/hr
SRb = 0.70 _ 0.01
1.10 1.20 1.30
OVERALL STOICHI0MFTRY - SR.
1.40
120

1.0 1.10 1.20 1.30
OVERALL ST0ICHI0METRY - SRT
1-40
Figure 5-12. Effect of excess air on 120 x 10^ Btu/hr DMB staged operation
with alternate host fuels.
-------
FUEL
<_n
. i
INI
600 -
500 -
CL
O-
>.
«-
¦o
s«
. o
100 -
O UTAH
~ ILLINOIS
UNSTAGED OPERATION
SRy = 1.20 t 0.01
80 90 100 110 120
50 60
90 100 110 120
FIRING RATE (10 Btu/hr)
FIRING RATE (10 Btu/hr)
/	Figure 5-13. Effect of firing rate on DMB 120 x 10 Btu/hr DMB
unstaged operation.	T
-------
a 45 percent turndown capability. The resulting flame was stable with a
shorter flame of 10-12 feet and N0X emission of 510 ppm. Burner operation at
firing rates below those referenced above was found to result in a deterio-
ration in flame stability performance.
A more detailed evaluation of reduced load operation was conducted with
Illinois coal and is summarized in Figures 5-14 and 5-15. Figure 5-14 shows
the effect of staging with reduced load. N0X is decreased by 30.2 percent as
staging 1s increased to SRg = 0.71. This Is significantly less than the full
load effect. However, at SRb = 0-7, NQX emissions are similar at both full
and reduced load operation with at about 340 ppm. CO emissions are slightly
higher at reduced load, approximately 50 ppm. The effect of excess air at
reduced load is slightly greater when compared to full load, with a 7 ppm/
percent excess air decrease in N0X and a 15 ppm increase in CO emission as
shown in Figure 5-15. At baseline staged conditions, the resulting flame was
stable and approximately 16-17 feet long.
5.4	60 x 106 Btu/hr DMB in Baseline LViS
The 60 x 1Q6 Btu/hr DMB was initially tested in the "improper" high
velocity outer secondary configuration. The so called short flame version
used a coal impeller similar in design to the device used in the pre-NSPS
Circular burner. A long flame arrangement, using a coal diffuser instead of
impeller, was also tested. The tests of the two initial 60 x 10® Btu/hr DMBs
were conducted in the baseline LWS.
5.4.1 Evaluation of Initial DMB with Impeller
The 60 x 106 Btu/hr DMB was initially tested with the coal impeller at
its baseline position in the coal nozzle as defined by B&W. A series of
tests were conducted with Utah Coal to determine the effect of register
adjustments on unstaged performance. For unstaged operation, the registers
effectively acted as flow control devices. Due to the increased pressure
drop through each secondary passage as the registers were closed, it was not
possible to operate at unstaged conditions with the registers at more than
. 5-22
-------
NOMINAL CONDITIONS:
FUEL -ILLINOIS
FIRING RATE =64.9 ! 1.4 x 10 Btu/hr
SRt = 1.20 t 0.01
cn
GO
700
600
500
E
£ 400
>>
&-
"O
^ 300
o
s«
o
200
X
o
100
Firing Rate
120x106 Btu
_L
X
_L
_L
X
0.60 0.70 0.80 0.90 1.00 1.10
BURNER ZONE ST0I6HI0METRY - SRg
0.60 0.70 0.80 0.90 1.00 1.10
BURNER ZONE ST0ICHI0METRY - SRg ['
Figure 5-14. Effect of staging on 120 x 10 Btu/hr DMB at reduced load.
-------
NOMINAL CONDITIONS:
FUEL - ILLINOIS	,
FIRING RATE = 65.4 1 1.8 x 10° Btu/hr
SRb = 0.71 t 0.01


600
1
i i i
120

I I 1 I


500
-
—
100
_
_

! I
O-
400

/D
S 80



i £•
XJ



>>
u
TJ


... 5-24
i
1 O
i ' O

X
o
z
i
300
200
i i
\
^ /
Firing Rate =
120 x 10K Btu/hr
OJ
S 60
O
<2;
O
O
40
-
~
120 x 106 Btu/hr

!
100
—
—
20
—
—


0
i
1 1 1
0

1 1 1 1
1.00 1.10 1.20 1.30 1.40	1.0 1.10 1.20 1.30 1.40
OVERALL STOICHIOMETRY - SRt	OVERALL STOICHIOMETRY - SRy
Figure 5-15. Effect of excess air on 120 x 1O^frtuThr DMETatT	/
reduced firing rate.	*
-------
40° and 30° closed on the inner and outer registers, respectively.
Figure 5-16 shows the effect of excess air on unstaged performance at these
burner settings. N0X emissions decreased by about 10 ppm/percent theoretical
air. CO emissions were minimum 65 ppm, at an overal1 stoichiometry (SRj) of
about 1.18. Either a reduction or an increase in total air from this point
resulted in an increase in CO emissions. The flame was stable throughout the
range tested with resulting flame lengths of about 14-16 feet.
An extensive test series was conducted to determine the optimum burner
settings for the 60 x 10® Btu/hr DMB at staged conditions. Figures 5-17 and
5-18 show the effects of register adjustment on emissions and flame length.
Optimum settings were 20° open counter clockwise and 10° open on the inner
and outer registers, respectively. These settings resulted in a reversed
flow pattern with counter-clockwise flow through the inner passage and
clockwise flow through the outer passage. The flame was stable throughout
the range of adjustment with lengths ranging from 16-20 feet.
The effect of staging by diverting air through the tertiary ports is
shown in Figure 5-19. To more fully characterize the effect of staging, two
different burner settings were used, since the above optimum register
settings precluded operation of the burner unstaged. By opening the inner
register to 40° (CCW) a wider range of staging was achievable. Generally,
the results are comparable with both settings. Staging the burner resulted
in a decrease of N0X emissions of 10 ppm/percent decrease in burner zone
stoichiometry (SRb) with no effect on CO emissions. At design point,
SR| =0.70 N0X was 390 ppm and CO was 60 ppm. The flame was stable, rooted
within the burner exit, with the main body of the flame about 16-17 feet in
length. However, the flame was not well-defined with combustion appearing to
occur throughout the entire length of the furnace with evidence of flame
licking the rear wall. Figure 5-20 summarizes the effects of excess air on
burner performance. N0X was reduced by approximately 8 ppm/percent excess
air. CO emissions were not significantly effected by excess air, remaining
stable at about 75 ppm down to an excess air level of 7 percent. Compared to
the 120 x 10® DMB performance, unburned carbon was found to be high and
ranged between 8 percent and 12 percent for optimum staged conditions. A
"5-25
-------
Nominal Conditions:
Fuel - Utah
Firing Rate = 58.6 - 1.1x10** Btu/hr
Q.
Q.

Register Position :
800 	r
700 \-
600
500
O
O
z
.x 400 -
300 -
Inner 40 Open CCW
Outer 30 Open
120
1.0 1.1 1.2 1.3 1.4
Overall Stoichlometry - SRT
100 -
&
£
T>
80 _
60
o
a*
o
o
o
40 -
20 -
1.0 1.1 1.2 1.3 1.4
Overall Stolchiometry - SRT
Figure 5-16. Effect of excess air on unstaged 60 x 106 Btu/hr DNIB performance.
-------
Nominal Conditions:
Fuel - Utah
Firing Rate =59.5 t I.TxlO6 Btu/hr
SRB = 0,70 + 0.01
Sfcf ¦ 1.21 i 0.02 _ ... - -
Outer Register (°0peri)
O 10
~ 20
- "800


7	1	1	1	1	
I 600
¦>

-
' 400
CM
o


«a
*«.
o
»
CM
O
(Pi
a
<_>
100
90|»
80T
70
60
50-
40.
30-
I 1 t I I
-i—i—i—i—r
Q-o^ u u n u
u u o u
o
__P	O
T—D
en
c
«
eu
E
L	Clockwise	> I
Closed
~>|«—
Open
CIosed
.yt-nner-Reg-tstei^; (°0pen)
Figure 5-1.7. Effect of inner register on 60 x 10 Btu/hr
staged performance.
DMB
5-27
-------
so o;
r» 600

2 200
T3
90
80
70
CM 60
o

o
o
50
40
30
0
j	L
Nominal Conditions;
Fuel - Utah
Firing Rate » 58.7 t 1.3xl0£
5R|
SR,
B
0.70 t 0-01
1.21 t 0.01
Inner Register ( Open]
o
20
ccw
a
40
ccw
a
50
ccw
Q
60
cw
0
30
cw
o
10
cw
- 24
H-
w 20
-C
•W
2? 16
«
E
-------
(_n
{i
PO
vo
Nominal Conditions:
Fuel - Utah	fi
Firing Rate =61.1 - 1.1 x 10 Btu/hr
SR,
= 1.20 + 0.01
Burner Settings ( Open)
Inner: Outer:
O 40 (CCW) 10
Q 20 (CCW) 10

700
1 1
1
1
i
140

1
1
1 1
I

600
-

f)
-
120
-



-
E
CL



r

,		






500
_
/

—
|ioo
_



_
¦t











•o











CVJ
o
400
j—|CJ


-

-



-
as
o
di
X
o
z
300
-
/


-
M
O
o
& 60
o
u

-a-o
-

200
~


-
40
-



-

100
-


-
20
-



-

0
1 1
1
i
i
0

1
1
• •
1
0.5 0.6 0.7 0.8 0.9 1.0
Burner Zone Stoichiometry - SRn
D
0.5 0.6 0.7. 0.8 0.9 1.0
Burner Zone Stoichiometry - SRn
D
Figure 5-19. Effect of staging on 60 x 106 Btu/hr DMB performance.
-------
ICO '¦
!o
Nominal Conditions:
Fuel - Utah
Firing Rate
59.0 - 0.95 x 106 Btu/hr
SR,
B
= 0.71 t 0.02
600
Register Position:
Inner - 20° Open CCW
Outer - M0° Open
400
-300
<=> 200
X
1.0 1.1 1.2 1.3 1.4
Overall Stoichiometry - SRy
cl 80
1.0 1.1 1.2 1.3 1.4
Overall Stoichiometry - SR^
Figure 5-20. Effect of excess air on 60 x 10 Btu/hr DMB performance.
-------
contributory factor here is believed to be the relatively cold furnace
environment.
The 60 x 10® Btu/hr DMB performance with two host coals, Comanche and
Wyodak, and with Illinois Coal was also evaluated. Figure 5-21 summarizes
the significant effect of staging with the Comanche Coal. Increasing staging
from SRg = 0.82 to SRg = 0.65 reduced N0X emissions by about 19 ppm/percent
theoretical air. At design point operation CSRg = 0.70 N0X was 420 ppm,
approximately 30 ppm higher than with Utah Coal. The CO emissions for the
Comanche Coal were low at 43 ppm throughout the range tested, significantly
1 ower than the corresponding emissions with Utah coal. The flame length was
shorter {13-14 feet), but lacked the intensity of the Utah flame.
Initial tests with the Wyodak Coal presented problems in flame stability
and operational difficulties. At optimum conditions, the flame was detached
about 10-12 feet from the burner exit. The flame could not be retracted with
the burner adjustments, available. A second attempt to fire the Wyodak Coal
was made at a later date, with the furnace heated by previous tests with
another coal. With these furnace conditions, it was possible to maintain a
flame stabilized within the burner exit. The reasons for this change in
performance are not clear, but are probably due to the characteristics of the
subbituminous Wyodak Coal and the furnace thermal environment. Decreased
primary air flow, resulting from increased mass flow of coal to achieve
design firing capacity limited operating range to an SRb of 0.65. .(A similar
but less severe problem was also encountered with the Comanche coal.) At
these conditions N0X and CO emissions were 445 and 49 ppm, respectively. The
flame was dull in color and approximately 14-15 feet in length.
Since, the two host sites utilize low sulfur coals and the baseline Utah
coal is also low in sulfur, high sulfur Illinois Coal was evaluated to
provide data which would be applicable to eastern U.S., boilers burning high
sulfur fuels. Figure 5-22 summarizes the effect of staging with the Illinois
coal. N0X emissions were reduced by 12 ppm/percent decrease in SRb- Optimum
conditions at an SRb of 0.65 resulted in N0X and CO emissions of 350 ppm and
-------
Nominal Conditions:
Fuel - Comanche
Firing Rate = 61.3 t 0.8 x 10^ Btu/hr
SRb = 1.20 t 0.02
700
<=> 300
*200
0.5 0.6 0.7 0.8 0.9
Burner Zone Stoichlometry - SR
Register Positions ( Open)
Inner:	Outer:
O 40° CCW
~ 20° CCW
10u
10°
140
B
0.5 0.6 0.7 0.8 0.9 1.0
Burner Zone Stoichlometry - SRg
Figure 5-21. Effect of staging on 60 x 10 Btu/hr DMB with Comanche coal
-------
Nominal Conditions:
Fuel - Illinois Coal
Firing Rate = 60.0 ± 0.9 x 10^ Btu/hr
SRt = 1.21 ± 0%
Reg. Position Open
INNER	OUTER
O 40° CCW	10°
~ 20 CCW
10L
E
a.
O.
~ 400 -
or
i
CO
CO
£
-o
o
4* 300 -
0.5 0.6 0.7 0.0 0.9 1.0
Burner Zone Stolchiometry - SRD
D
0.5 0.6 0.7 0.8 0,9 1.0
Burner Zone Stoichiometry - SRg
/Figure 5-22. Effect of staging on 60 x 10 Btu/hr with Ill_inois_.coal._
-------
61 ppm, respectively. The flame was stable throughout the range tested with
length of approximately 19-21 feet.
The effect of excess air on the alternate fuels is summarized in
Figure 5-23. Excess air had a similar effect on Illinois and Comanche Coal
with a reduction in N0X emissions of approximately 9 ppm/percent excess air.
N0X emissions for Wyodak were about 80 ppm higher than with Illinois, but as
excess air was reduced, the difference in emissions was less than 15 ppm.
The highest CO emissions of 65 ppm resulted for the Illinois Coal, with the
characteristic CO "knee" at SRi = 1.13. The host coals showed considerably
lower CO levels, typically less than 50 ppm, and good carbon burnout.
Key results at optimum conditions for the impeller equipped 60 x 10®
Btu/hr DMB are summarized in Table 5-1.
5.412 Evaluation of Initial DMB with Coal Piffuser
The long flame configuration of the 60 x 10® Btu/hr DMB was achieved by
installing a coal diffuser back in the coal pipe near the coal inlet of the
burner in place of the impeller used for the short flame design. Long flame
developmental DMBs were found to produce lower N0X emissions than DMBs which
produced shorter flames. However, long flame burner designs could only be
used in installations which could accommodate the long flames, such as
boilers which utilize opposed-fired B&W Dual Register Burners (DRBs). A
brief series of tests was conducted with Utah coal to optimize burner
performance with regard to emissions and flame characteristics.
The effect of register adjustment on staged burner performance is
summarized in Figure 5-24. The flame was over 22 feet long, the firing depth
of LWS, and 6-8 feet wide throughout the entire range of register adjustment.
The tests of the short flame DMB 60 indicated that flame length was most
sensitive to outer register position. The flame could be shortened by
closing the outer register and this decreased the degree of swirl. With this
long flame configuration, however, even register settings of only 20° and 10°
open produced excessive flame length. Attempts to shorten the flame by
5-34
-------
Nominal Conditions:
Tiring Rate = 60.0 * 1.2 x 10 Btu/hr
SRg = 0.66 t 0.02
FUEL;
U Uyodak
CD Comanche
A Illinois
600
500 -
1.1	1.2	K3
Overall Stolchlometry - SR^

¦*» 60
1.1	1.2 1.3
Overall Stolchlometry - SR^
Figure 5-23. Summary of effect of excess air on 60 x106 Btu/hr DMB
staged performance.
-------
TABLE 5-1. SUMMARY OF THE 60 x 106 Btu/hr DMB.SHORT FLAME BURNER OPTIMUM CONDITIONS
Test
No.
Fuel
Firing Rate
(106 Btu/hr)
srb
srt
N0X, ppm
(0% 02)
CO, ppm
(0% 02)
Flame
Length (ft)
3.10
Utah
61.1
0.69
1.19
390
60
16-17
9.02
Wyodak
59.4
0.66
1.21
449
49
13-14
4.02
Comanche
61.1
0.68
1.23
420
43
12-14
6.07
111i noi s
59.8
0.65
1.21
348
61
19-21
Note: OPTIMUM BURNER CONDITIONS:
Inner Register - 20° Open CCW
Outer Register - 10° Open CW
Spreader - 0
-------
Nominal Conditions:
FUEL - Utah
FIRING RATE - 60.4
Outer Register Position
O 10° Open CW
m 20° Open CW
Btu/hr
a 3oc- o
i__i
-------
creating counter flow between the two secondary air passages were not
successful. Opening the inner register to a position of 40° open ore more
resulted in unstable flames of even greater length. The best overall
performance was achieved with the inner register set to 20° open clockwise
and the outer register at 10° open. However, even these settings produced a
flame which impinged on the rear wall of the LWS.
To avoid potential damage to the facility, the tests were abbreviated
and only Utah coal was evaluated. The effect of staging on emissions from
the long flame 60 x 10® Btu/hr DMB is shown in Figure 5-25. N0X emissions
decreased by 13 ppm/percent theoretical air as staging was increased from
SRg = 0.86 to SRg = 0.61. The CO emissions increased by approximately 30 ppm
over the same range of staging. The flame was stable throughout the range
tested with the minimum length of 21-22 feet at SR® = 0.86. At design point
conditions SRg = 0.70 N0X and CO emissions were 230 ppm and 63 ppm
respectively. This long flame design resulted in a 41 percent reduction in
N0X emissions from the short flame configuration with no significant change
in CO emissions.
Figure 5-26 shows the effect of excess air on the long flame DMB 60.
N0X emissions decreased at about 6 ppm/percent excess air. The CO emissions
were minimum at the design point SRj = 1.2, at about 62 ppm.
5.5	60 x IP6 Btu/hr DMB in Insulated LWS
5.5.1 Characterization of Initial DMB with Impeller
Additional insulation was installed above the baseline refractory up to
the nose of the LWS to produce a thermal environment at 60 x 10® Btu/hr input
similar to that produced in the baseline configuration at full load,
120 x 10® Btu/hr. After the installation of the ceramic board insulation,
the short flame 60 x 10® Btu/hr DMB was retested to evaluate the effects of
thermal environment on its performance. The initial tests were conducted
with Utah coal . The burner conditions determined to be optimum in the
baseline LWS were maintained during these tests to provide a direct
5-38
-------
Nominal Conditions:
Fuel - Utah	.	fi
Firing Rate = 60.4 ~ 0.5 x 10 Btu/hr
SR,
1.20 t; 0.01
Register Position ( open)
Inner - 20 CIJ
Outer - 10 CH
700
600
5.0 i
.60 .70 "780 75G TTO
Burner Zone Stoichiometry - SRR
.50 .60 .70 .80 .90 1.00
Burner Zone Stoichiometry - SRg
-Figure 5-25. Effect of staging on diffuser equipped DMB performance.
-------
NOMINAL CONDITIONS:
FUEL - UTAH	,
FIRING RATE = 60.7 I 0.55 x 10° Btu/hr
SBb = 0.69
0.01
REGISTER POSITIONS (° Open)
INNER - 20 CW
OUTER - 10 CW
60Q
CJ1 I
ii r
o
'I
>,
L.
TD
CM
O
o
40 _
1.00
1.10
1.20
1.30
1.40
Q
Q.
>>
J-
"O
CvJ
O
o
o
u
OVERALL ST0ICHI0METRY - SRn
1.00 1.10 1.20 1.30
OVERALL ST0IGHI0METRY - SRT
1.40
Figure 5-26. Effect of excess air on diffuser equipped DMB performances
-------
comparison. An array of six Type K thermocouples were utilized in the exit
of the LWS to evaluate temperature differences. In addition, a suction
pyrometer was positioned adjacent to one of the thermocouples to provide a
calibration of radiation induced errors in temperature measurement. Another
suction pyrometer was installed in the middle of the furnace about 8 feet
above the burner center!ine. Table 5-2 summarizes these temperature
measurements. Variations in these temperatures are generally due to the
change in ash and slag build up over time in the furnace. The insulation
increased furnace temperatures on the average approximately 200°F.
The effect of staging and excess air on emissions from the short flame
60 x 10® Btu/hr DMB in the insulated LWS are shown in Figures 5-27 and 5-28,
respectively. N0X decreased at a rate of about 14 ppm/percent theoretical
air as staging increased, compared to about 10 ppm/percent theoretical air in
the baseline LWS. Overall, N0X emissions with the Utah coal were higher in
the insulated furnace. At a nominal burner zone stoichiometry of 0.7, N0X
emissions were 540 ppm compared to only 390 ppm in the baseline LWS. CO
emissions, an indication of combustion efficiency, were lower in the
insulated furnace, 51 ppm compared to 60 ppm and measurements indicated that
carbon in ash was considerably reduced (to 2.5 percent) compared to the cold
furnace. The flame was stable over the range of staging tested, with lengths
of 20-22 feet. N0X emissions were much more sensitive to excess air in the
insulated furnace than for the baseline configuration, with rates of change
in N0X of 14 and 8 ppm/percent excess air, respectively. This data suggests
the enhancement of thermal N0X formation and the resulting increased
sensitivity to combustion stoichiometry.
The 60 x 10® Btu/hr DMB performance with two host coals, Comanche and
Wyodak, and with Illinois Coal was also* evaluated in the insulated LWS.
Figure 5-29 summarizes the effect of staging of the alternate fuels.
Increasing staging yielded similar results for each coal with N0X increasing
at approximately 10 ppm/percent increase in theoretical air. At SRg = 0.7,
N0X was 446 ppm for Comanche which is about 25 ppm higher than in the
baseline furnace. There was no significant change in CO emissions at about
38 ppm. The flame length was about 16-18 feet over the range tested and
5-41
-------
TABLE 5-2. SUMMARY OF LWS FURNACE TEMPERATURES DURING 60 x 106 BTU/HR
DMB TESTS
Furnace	Exit Temperature Exit Suction OFA Suction
Configuration (Bare Type K), °F Pyrometer, °F Pyrometer °F
Baseline	1374-1547	1484-1640	1818
Insulated	1590-1682	1732-1819	2095
-------
NOMINAL CONDITIONS:
FUEL - UTAH
FIRING RATE = 59.4 + 0.90 x 106 Btu/hr
SRt = 1.20 t 0.01
REGISTER POSITION (° OPEN)
INNER - 20 CCU
OUTER - 10 CM
.50	.60 .70	.80	.90
BURNER ZONE ST01CH10METRY - SRB
CD OOP O
On O
.50	.60 TUT
BURNER ZONE STOICHIOMETRY
^6
	Figure 5-27. Effect of staging on 60 x 10" Btu/hr DMB
performance in insulated LWS.
-------
NOMINAL CONDITIONS:
FUEL = UTAH
FIRING RATE = 62.2 1 23 x 106 Btu/hr
SRb = 0.65 t 0.03
REGISTER POSITION (° OPEN)
INNER - 20 CCU
OUTER - 10 CH
1.00 1.10 1.20 1.30
Overall Stolchlometry - SRf
1.40
120 -
100 -
"1.00 I.10 1.20 1.30
Overall Stolchlometry - SRj
1.40
		 					_ _ _ _ K_		¦	 	
Figure 5-28. Effect of excess air on 60 x 10	Btu/hr' DMB performance in insu-
.1ated LWS.	'	^
-------
cn
i
-P»
cn
NOMINAL CONDITIONS:
FUEL
O COMANCHE
~ ILLINOIS
A WYODAK
FIRING RATE = 60.9 t 1.7x 106 Btu/hr
SR,
1.20 + 0.01
/
>< 20C _
REGISTER POSITION (° OPEN)
INNER - 20 CH OUTER - CVI
14G
CtBD
Figure 5-29.
-------
again lacked the intensity of the Utah flame. As in the previous tests with
Wyodak coal, decreased primary air flow, resulting from increased mass flow
of coal to achieve design firing capacity, limited the operating range of
staging to SRg = 0.65. CO emissions were slightly lower with 43 ppm. The
flame was stable, but dull in color with a length of about 17-20 feet. High
sulfur Illinois coal was evaluated to provide data which would be applicable
to U.S. boilers burning eastern high sulfur fuels. Optimum conditions at
SRg = 0.65 resulted in N0X emissions of about 456 ppm which is 100 ppm higher
than in the uninsulated configuration. CO emissions were substantially lower
at 43 ppm compared to 61 ppm. The flame was stable throughout the range
tested with lengths of about 20-22 feet.
The effect of excess air on 60 x 10^ Btu/hr DMB performance with
Comanche and Illinois coals is shown in Figure 5-30. Overall combustion
stoichiometry had a similar effect on N0X for the two coals, decreasing
8 ppm/percent excess air. This slope is similar to that*measured in the
baseline furnace. Both fuels resulted in similar N0X emissions for the
excess air range tested, with slightly lower CO produced by the Comanche
coal.
The effect of excess air on the 60 x 106 Btu/hr DMB in the insulated LWS
with Wyodak Coal is summarized in Figure 5-31. N0X emissions decreased
8 ppm/percent excess air over the range tested. The flame was stable, but
with slight detachment of about 6-12 inches from the burner throat and was
about 16-18 feet in length. At baseline conditions SRg = 0.70 and SRj = 1.20
N0X emissions were approximately 500 ppm. Also" shown is the effect of excess
air in the uninsulated LWS. N0X emissions are generally 50 ppm lower in the
uninsulated LWS over the range tested. CO emissions were not significantly
different {5 ppm) between the two furnace thermal configurations.
5.5.2 Evaluation of Modified DMB
The initial 60 x 106 Btu/hr DMB had been incorrectly scaled, resulting
in very high secondary air velocities. The modified DMB evaluated
incorporated design parameters which matched the 120 x 106 Btu/hr DMB. In
. 5-4.6. ¦
-------
NOMINAL CONDITIONS:
ui
i
-P»
o
~
Comanche
Illinois
Firing Rate = 60.9 * 1.7 x 10® Btu/hr
SRb = 0.65 t 0.008
Register Position (° open)
Inner - 20
Outer - 10
"T"
1.00 1.10 1.20 1.30
Overall Stolchiometry - SRj
<&
O
CNo "* -q—-Q-O
1.00 l.l'O 1.20 1.30
Overall Stolchiometry - SRj
r
[	Figure 5-30. Effect of excess air on _60 x 10 Btu/hr DMB with
1	alternate f uel.s _ arid insulated _LWS_
-------
NOMINAL CONDITIONS:
FUEL WYODAK
FIRING RATE = 59.6 t 0.50 x 106 Btu/hr
SRb = 0.69 t 0.01
ui
i
oo
600
500 _
I.
CL
400 -
u
*o
o
200 -
100 _
11HOT"

"COLD" LWS
X
_L
1.0 1.10 1.20 1.30
OVERALL STOICHIOMETRY (SRj)
1.40
120
100
I 80
>>
t-
T3
CM
o
o
o
o
60
40
20
"COLD" LWS
"HOT" LWS
_L
I
X
1.0 1.10 1.20 1.30
OVERALL STOICHIOMETRY ( SRT)
1.40
,6:
Figure 5-31. Summary of 60 x 10 Btu/hr DMB burner performance[ with_Wyo.d.ak_coal.	/
-------
addition to adjusted burner velocities, the optimum coal spreader from the
120 x 1Q6 Btu/hr DMB was scaled down for this redesigned 60 x 10® Btu/hr DMB.
Thus, the redesigned small DMB matched the larger scale 120 x 106 Btu/hr DMB.
Parametric optimization of the redesigned 60 x 106 Btu/hr DMB was
conducted using Utah coal fired at nominal full load. The burner was
operated at nominal design point air flows throughout these optimization
tests, with primary stoichiometry of 0.20, burner zone stoichiometry of 0.70,
and overall stoichiometry of 1.20. The secondary air was divided equally
between the inner and outer passages. Since the DMB design had been
optimized with the larger-scale 120 x 106 DMB, the parametric optimization
for this small DMB was limited to the following adjustable burner parameters:
•	Coal spreader position
•	Outer register vane position
•	Inner spin vane position
•	Secondary air distribution
The effect of these parameters are briefly described below;
Coal spreader position. The reference or zero, position of the spreader was
defined with the leading edge of spreader flush with the end of the coal
nozzle. With the coal spreader advanced 2 inches beyond the coal nozzle into
the furnace, N0X emissions were measured as 372 ppm (at 0 percent 02 dry).
As the coal spreader was retracted to the zero position N0X emissions
£
decreased to 348 ppm, and then increased to 382 ppm as the spreader was
further retracted to a position 3 inches inside the coal nozzle. The flame
length varied between 18 and 20 feet for this entire range of adjustment.
Outer register vane position. The outer register vanes were varied over a
range from 10° open to 50° open in a clockwise direction. N0X emissions were
lowest (336 ppm) at the more open, lower swirl position of 50° with a flame
length of 20-22 feet. At a position of 10°, which matches the optimum
5-49
-------
setting for the 120 x 10® Btu/hr DMB, N0X emissions were about 367 ppm with
18-20 foot flames.
Inner spin vane position. The spin vanes were adjusted from the 25°
clockwise to the 20° counter clockwise positions. Over this range of
adjustment, N0X emissions varied from 322 to 459 ppm with flame lengths from
12 to 21 feet. The best performance was achieved at 55° clockwise with
372 ppm and flames about 13 feet long. The corresponding outer register
setting was 10°CW.
Secondary air distribution. Secondary air distribution was varied from 10
percent to 60 percent biased toward the inner passage.
From these parametric tests, the burner parameters which yielded the
best performance for the 60 x 10® Btu/hr DMB were:
The staged performance of the 60 x 106 Btu/hr DMB was characterized at
effect of staging for each coal is shown in Figure 5-32, N0X emissions were
lowest for the Utah coal. At the design point burner zone stoichiometry of
0.70, N0X emissions were 340, 440, and 525 ppm for Utah, Comanche, and Wyodak
coals, respectively. The carbon content of fly ash samples were less than
0.5 percent for the two subbi tumi nous coals, Comanche and Wyodak, and
1.5 percent for the Utah coal at design point conditions.
With many low-NOx burners, an increase in carbon-in-ash is normally
expected as staged combustion is applied. For the 60 x 10® Btu/hr DMB
burner, however, the results of Figure 5-32 indicate that there is no
deterioration in carbon burnout for the two subbi tumi nous coals, and only a
marginal charge for Utah coal, even though substantial reductions in N0X
Coal spreader position
Inner spin vanes
Outer register
Secondary air distribution
= 0 inches
= 55° CW
= 10° CW
= 50% inner/ 50% outer
the above listed optimum settings for Utah, Wyodak, and Comanche coals. The
5-50:
-------
Nominal Conditions:
CJ1
I
un
FiringJ?ate = 60 x 10 Btu/hr
SR-,
Fuel:j O i Utah
~
A
Wyodak
Comanche
1.20
SRp 1 = 0.20
- 400
0.50 0.60 0.70 b.80 0.90 1.00
0 A
SR r
0.50 0.60 0.70 0.80 0.90 1.00
SRd
Figure 5-32. Effect of staging on modified 60 x 10 Btu/hr DMB performance.
-------
emissions are achieved. This is believed to be due in part to the low firing
rate which provides long residence times for burnout to occur. Previous
results for the 120 x 10® Btu/hr DMB have shown a much stronger effect of
staging on carbon burnout for the Utah coal. Burnout with the subbituminous
coals was found to be consistently good and relatively insensitive to
staging. This may be associated with the higher reactivity of these fuels.
The effect of excess air at staged conditions, with a burner zone
stoichiometry of 0.70, is shown in Figure 5-33 for all three coals. Again,
N0X emissions were lowest for Utah coal and highest for the Wyodak coal over
the range of excess air evaluated. Again, unburned carbon levels were
exceptionally low, 1 ess the 0.5 percent carbon, for the subbituminous coals.
Combustion efficiency was more sensitive for the small DMB with the Utah
coal, with up to about 3.2 percent carbon in the fly ash at an overall
stoichiometry of 1.10.
The 60 x 10® Btu/hr DMB was also evaluated briefly unstaged with all
three coals. The results are shown in Figure 5-34. Flame lengths under
these unstaged conditions ranged from 10 to 13 feet for the subbituminous
coals and about 17-18 feet for the Utah coal. N0X emissions were lowest for
the Utah coal with 775 ppm at an overall stoichiometry of 1.20. N0X
emissions at the same excess air level were 850 and 900 ppm for the Wyodak
and Comanche coals, respectively. As with staged conditions, carbon in fly
ash was less than 0.5 percent for the subbituminous coals. Again, the
combustion efficiency of the DMB was more sensitive when firing Utah coal,
although the level of carbon in the fly ash was still very low (less than
2.5	percent).
5.6	Discussion of Results and Extrapolation to Full Scale
The preceding sections have presented N0X emission and burner
performance data for the four major burner designs which were the subject of
this program. While this discussion has provided information on individual
burner performance the overall burner test program was structured to provide
data to address the performance scale-up issues of:
5-52
-------
o
LT>
1
i t r ¦
i i 1
o
-

1
vo {
X


¦°



'
i l i
1 1 1
O
o
CO
o
CM
CC
OO
o
O
o
ln
o
ir>
o
Lft
o
LT)
«3-
CO
no
CM
CVJ

—
o
o
o
(% "V*) 'MStf J ui- uoqueo
S-.
.c
3
•W
CO
o
CO
o

c
VO

o


•r-

-C
+->
II
rO


¦M
"O


c
a>

o
o
-M
fO
a:
O



 •


i—
E
j-
0)
o
¦r—
3
z
LU
Lu
a:
oo
(njdd) /Up ,zo %0 0 *0N
5-53
-------
Nominal Conditions:
Firing Rate = 60 x 10 Btu/hr
Fuel: O Utah
Q Wyodak
A Comanche
SRp = 0.20
Unstaged
1000
O.
CL
>»
u
x»
CO
o
o

&
Figure 5-34. Effect of excess air on modified 60 x 10 Btu/hr DMB performance
under unstaged conditions.
-------
•	Thermal environment
•	Burner design/capacity scaling-
•	Single burner-to-multiple burner installations
The following sections will summarize the relative performance data and
provide discussion on aspects related to scale-up and extrapolation to full-
scale boiler systems.
5.6.1 Relative Burner Performance and Effects of Coal Type
The evaluation of a full-scale 120 x 106 Btu/hr Distributed Mixing
Burner was a key element of this demonstration program. For burners of this
size the LWS test furnace imposes some significant constraints on flame shape
and length, since 1ow-N0x burners, like the DMB, rely on controlled, delayed
mixing of the fuel with air. This delay generally produces a long flame
which may cause operational problems in a boiler. Although equipped with
adjustable inner and outer secondary air parameters as well as the tertiary
air ports, the dominant factor in determining ultimate performance (N0X.
flame length) was the coal injector configuration. Iterative modifications
were made to the coal injector to yield the optimum performance for the LWS.
There was a direct tradeoff between N0X emissions and flame length.
The final design selected resulted in unstaged flames about 16 ft long.
At staged conditions (SRg = 0.70), the flame length increased to
approximately 22 ft. The optimum configuration for the 120 x 106 Btu/hr DMB
was determined to be;
•	Spreader design = 4-inch support pipe with four 8-inch blades at a
30° angle from axial,
•	Burner settings: Inner spin vanes = 35° open CW
Outer register = 10° open CW
Spreader portion = 3 in. retracted
Inner/outer secondary air distribution = 50/50
5-55.
-------
N0X emissions for the DMB at these optimum settings at nominal full load
conditions with a burner zone stoichiometry of 0.70 and 20 percent excess air
were 300, 340, 298, and 273 ppm for Utah, Illinois, Comanche and Wyodak
coals, respectively.
Performance data for the four main burner types are summarized in
Table 5-3, for operation on Utah coal, which was the design coal for the DMB
burners. In this table the half scale DMB data represents burner design
parameters scaled from the 120 x 10® Btu/hr version, and test conditions for
the two 60 x 10® Btu/hr burners are based on an insulated furnace where
temperature levels are comparable to those at the higher load. The
conditions listed represent stable operation where flame characteristics and
carbon burnout were acceptable.
Further information comparing burner N0X emission performance over a
range of excess air levels and for different coals is presented in Figure
5-35. The data in this figure indicate that all burners tend to respond
differently to these test parameters. Both of the 60 x 106 Btu/hr burners
{the DRB and the DMB) show a strong sensitivity of N0X emissions to Increased
excess air {100-150 ppm per % O2), compared to the 120 x 10® Btu/hr burners
for which this trend is less marked (50 ppm per % 02). Although similar
trends have been observed previously with comparable burner designs, other
LWS data has suggested that fully optimized low-N0x burners are relatively
insensitive to many operational parameters. One reason for this behavior may
be' due to the comparatively high local temperatures of the insulated furnace
configuration, and to the higher intrinsic mixing rates associated with the
smaller burner dimensions.
Of particular interest in Figure 5-35 is the performance of the
different burners with the different coal types. For the 60 x 10® Btu/hr DMB
N0X emissions with Wyodak coal are considerably higher than those achieved
with the Utah coal. For the DRB burner the reverse is true, while
comparatively small differences between coal types are observed for the 120 x
10® DMB. For" the 60 x 10® Btu/hr DMB burners, particularly In the colder
baseline furnace configuration, some difficulties were experienced in
5-56
-------
TABLE 5-3. COMPARISON OF BURNER PERFORMANCE IN THE
LWS FIRING UTAH COAL {SRj = 1.20)
DMB
Full Scale
Half Scale
DRB
Half Scale
'"Circular.
Full Scale
Firing Rate (10® Btu/hr)
SRB
FEGT {°F)
N0X (ppm @ 0% 02)
Flame Length (ft)
Carbon in Ash (wt %)
120
0.70
1792
300
22
5.5
60 .
0.70
1776
350
18
1.5
60
0.70
1776
390
18
0.9
120
1.20
r
1828
380
>22
5.2
FEGT = Furnace3Exit
Gas Temperature
• 5.^57..
-------
Coal Type:
Utah
Illinois
Comanche
Wyodak
Circular
120 x 106 Btu/hr
600 -

E
Q.
Q.


>>
S-
o
400
lUl
I 1
CM
o

;cn
00
>5
O


NO
X
200
A,-
1.1 1.2 1.3
J DRB
60 x 106 Btu/hr
DMB
:120 x 106 Btu/hr
1.1 1.2 1.3	1.1 1.2 1.3
.Overall Stoichiometry—SRj
DMB
60 x 10® Btu/hr
'/
1.1 1.2 1.3
Figure 5-35. Effects of burner design and coal type on NO emissions.
X
-------
achieving stable ignition for the Wyodak and Comanche coals under staged
combustion conditions. This may have limited the achievement of lower NCX
emissions for these coals, and suggests that further design changes might be
necessary to accommodate the lower heating value, higher moisture
subbituminous coals.
In spite of difficulties in achieving acceptable flame stability under
some conditions for Wyodak and Comanche coals, carbon burnout was generally
good for these fuels, and commensurate with the expected high reactivity.
For all burner designs, staging conditions, and under cold furnace
conditions, the measured carbon in ash was never found to exceed one percent.
This is a further indication that lower N0X emissions should be possible
with the subbi tumi nous coals without compromising overall combustion
performance.
For the two high volatile bituminous coals (Utah and Illinois) carbon
burnout was found to be more sensitive to combustion staging and to firing
rate. Carbon in ash values were generally found to increase as staging was
applied, and were higher for the higher capacity burners. However, for most
major operating conditions carbon in ash values were considered to be
acceptable (5 percent or below), and only deteriorated significantly for
operation at 60 x 10® Btu/hr in the baseline furnace configuration.
Additionally both of the high capacity burners and both of the low capacity
burners tended to yield similar burnout performance with the bituminous
coals. These observations would tend to suggest that the measured carbon
burnout values are more dependent upon coal type and furnace conditions than
upon burner design.
5.6.2 Extrapolation of N0X Emission Data
The correlation of N0X emission data between different firing conditions
in the test furnace, and the extrapolation of test data to full-scale systems
is of particular interest the overall goals of this program. While
fundamentally based computational tools are under development for the
prediction and extrapolation of N0X formation as influenced by burner and
:"5-59 ¦
-------
furnace parameters, most practical approaches are based on empirical
relationships. In fact, all major boiler manufacturers have used a burner
area heat release parameter at. some stage in an attempt to correlate N0X
emissions from full-scale boilers, and in some cases to try and relate/
extrapolate the results of burner tests in small-scale furnaces. Burner area
heat release rate 1s to some extent a measure or indicator of temperature
levels in the flame region. It is also indirectly related to the volumetric
heat release rate in the flame zone because most furnaces do not differ
greatly in geometry in the burner region. High volumetric heat release rates
reflect high air/fuel mixing and should thus correlate with N0X emissions.
Each of the manufacturers—Babcock & Wilcox, Foster Wheeler, and Riley
Stoker--has developed a correlation based on their own definition of burner
area heat release rate. The correlations are similar but not strictly
comparable because of differences in the methods of defining the cooled
surface area in the burner region. For purposes of comparison, N0X emission
data available to EER for boilers and large-scale burner tests have been
correlated using the B&W definition of cooled surface in the burner region
(HA/SC). This considers the burner area cooled surface to be the four sides'
and bottom of a cuboid box with the same width and depth as the furnace and
with a height extending half a burner spacing above and below the upper and
lower burner rows, respectively.
The use of the HA/SC parameter as defined above is illustrated in Figure 5-36
for N0X emission data from a wide variety of wall-fired boilers. A wide
range of boiler and burner designs, unit size and firing patterns are
incorporated into the data base. The data is, however, restricted to boilers
firing high volatile bituminous coals. Although there is considerable
scatter in these data points, an upper limit in N0X emission can be
identified, which increases with increasing burner zone heat loading. This
upper limit is indicated by a trend line for which the slope (0.8 x HA/SC) is
in good agreement with the correlation developed by B&W. Inspection of the
data which lie close to the upper limit shows these points to be associated
with boilers of pre-NSPS design in which high-turbulence burners are
employed. On Figure 5-36 the N0X emission data which lie below the upper
5-60
-------

1400
1200
1000
Q.
800
CM
O
w
O 600
400
200
0 Comanche ( no,, 0FA)
0 Wyodak
Upper Limit
Trend Line
.~^'^wLow-NOx
Extrapolation
P*"
0
200
400
600
800
HA/SC - Burner Zone Heat Release (103 Btu/hr, ft2)
1000
Figure 5-36. Correlation of boiler N0X emissions.
-------
limit line represent boilers with alternate burner designs, or situations
where different N0X control strategies have been applied. At the lower
levels of N0X emission the boilers have been retrofitted with low-NOx burners
and optimized for maximum N0X control. As might be expected, the HA/SC
parameter does not provide a universal correlation with N0X emissions, since
it fails to take into account the important effects of variables such as
burner design and coal type. Indeed the burner zone heat release parameter
was developed strictly to provide a means of extrapolating the performance of
a given burner design into an alternate firing situation. Also shown on
Figure 5-36 are baseline field data obtained from the two test boilers
(Wyodak and Comanche), indicating their N0X emission in relation to burner
zone heat release.
In order to incorporate experimental data into this kind of analysis,
the HA/SC correlation has been extended to include data obtained in the Large
Watertube Simulator furnace (LWS). In developing the HA/SC correlation there
is a problem in correctly representing the burner zone area for experimental
facilities. This is because small-scale furnaces have a comparatively high
surface-to-volume ratio, and consequently some degree of insulation on the
walls is required to more correctly match the thermal environment of full-
size boilers. Indeed, adding or removing refractory on the furnace walls has
been used in this program as a means of changing the thermal environment to
experimental ly evaluate the effect of this parameter. In order to overcome
this difficulty an "effective" burner zone cooled surface area has been
defined depending upon firing rate and extent of refractory insulation. The
effective HA/SC is determined by comparing maximum flame temperatures in the
combustion zone with corresponding temperature levels in full-scale boilers,
using heat transfer computer codes. Using these results, a correlation was
obtained between the maximum predicted flame temperature and the burner area
heat release rate calculated using the B&W definition. Using this
correlation and similar predictions of gas temperatures in the experimental
furnaces, effective burner belt cooling areas were obtained.
In Figure 5-37 selected experimental data are presented on the basis of
HA/SC defined in the above manner for the appropriate firing configurations.
5-62
-------
1000 .
800 -
E
o.
Q-
t? 600
-o
CM
> o
en'
i , **
OY O
CO
* 400 -
o
X
o
200 _
100
±
X
200	300	400
Effective HA/SC (103 Btu/hr ft2)
Oj 120 x 10 Btu/hr Circular
Burner
|0 120 x 106 Btu/hr DMB
60 x 10 Btu/hr DRB
N' 60 X 10 Btu/hr DMB
_L
500
SOLID SYMBOLS
BOILER DATA
OPEN SYMBOLS -
FURNACE DATA
HALF OPEN SYMBOLS
REDUCED SCALE
BURNERS
600
Figure 5-37. Correlation of experimental furnace data.
-------
The experimental data (open symbols) are for selected single full-scale
burners where corresponding boiler data (solid symbols) is available for the
corresponding data sets, where again it can be seen that a reasonably
consistent correlation exists. The use of a correlation of this type is,
however, believed to be conditional upon a number of factors, namely:
a)	The experimental data is derived from full-scale or near full-scale
burners.
b)	The same burner design and operating conditions are used both in
the experimental furnace and in the field.
c)	The same coal type is used at all scales.
d)	Experimental data is restricted to acceptable firing conditions;
i.e., where the flame fits within the confines of the furnace;
unburned carbon levels are low; the burner operates satisfactorily
over the required turn-down range.
Condition d) above is an obvious one, in that the extrapolation can be based
only upon operating conditions which are acceptable in the final application.
In this regard flame dimensions are an important consideration for Tow-NOx
burners, since , the flame must be constrained within the space available in
the boiler. This is related to condition a) which restricts the
extrapolation essentially to full-scale burners. If reduced scale burners
are used in the experimental evaluation, then another variable is introduced,
that of the effect of burner size on N0X emissions. The relationship between
burner design, burner size, the burner scaling approach employed, and NOx
emissions, is not well understood and what little information is available is
poorly documented.
Effective HA/SC values as defined above have been calculated for the
main test conditions of this program. These are;
• Baseline furnace, 60 x 10® Btu/hr: HA/SC = 89 x 103 Btu/hr-ft^.
5-64
-------
•	Baseline furnace, 120 x 106 Btu/hr: HA/SC = 178 x 103 Btu/hr*ft2.
•	Insulated furnace, 60 x 106 Btu/hr: HA/SC = 211 x 103 Btu/hr-ft2.
Using these values the test data for the four burner designs firing on Utah
coal are presented in Figure 5-37 to illustrate the relationship to previous
experimental N0X data.
The major issue addressed by the HA/SC parameter is the effect of
thermal environment or firing density on the emission of NQX from a burner of
fixed design. In this program, this parameter was investigated as an
independent variable by varying the insulation level in the IWS furnace. The
initial 60 x 10® Btu/hr DMB configuration was tested in the LWS with two
different insulation installations. The two levels of insulation yielded two
distinct thermal environments that can be characterized by the furnace exit
gas temperature. The LWS in its basic insulation scheme had an average exit
temperature of 1562°F when fired at 60x 106 Btu/hr. With additional
insulation, the exit temperature was increased to 1776°F. The level of
insulation was designed so that the exit temperature when fired at 60 x 106
Btu/hr would match the exit temperature achieved in the baseline LWS when
fired at 120 x 10® Btu/hr. As shown previously this insulation objective was
achieved successfully (within 35°F). The effect of thermal environment on
burner performance, specifically N0X emissions, is shown in Figure 5-38. The
effect varies with coal type. There was only a small effect on emissions
from either subbituminous coal, while for the high volatile bituminous coals,
N0X emissions were about 100 ppm higher in the insulated furnace.
The effect of thermal environment on N0X emissions is further
illustrated in Figure 5-39 where, for the short flame 60 x 10® Btu/hr DMB,
HA/SC is used as parameter. Data for nominal burner operating conditions of
SRg =0.7 and SRj = 1.2 is used throughout. Here it can again be seen that
for the Utah and Illinois coals there is a constant relationship between HA/
SC and N0X emission for the same burner fired under different thermal
conditions. For these coals, the increase in N0X emission is approximately
1.0 x HA/SC and is comparable to trends derived earlier and presented in
Figures 5-36 and 5-37. However, for the two subbituminous coals, under cold
5-65
-------
Load :
srt :
NOTE:
60 x 10° Btu/hr
1.20
Shaded symbols in LWS w/insulation
UTAH
cn
, i
CTV
£? 300
5 *00"
0.5 0.6 0.7 0.8 0.9 0
SRb
ILLINOIS
.5 0.6 0.7 0.8 0.9 0.5 0.6 0.7 0.8 0.9 0.5 0.6 0.7 0.8 0.9
SR.
SR„
Figure 5-38. Effect of thermal environment on NO emissions from initial
		short flame 60 x 10^ Btu/hr DMB.	
> • 	,!
-------
600
c_n
i
cr>
500
Q.
CL
>>
i-
O
400
300
o
9* 200
100
DMB Burners
SRg = 0.7
SRt = 1.2
I
I
©
I
o

a
i
rm
B
i
i
w	I
0 Long-Flame 60 x 10® Btu/hr DMB
O Short-Flame 60 x 106 Btu/hr DMB
A Modified 60 x 10® Btu/hr DMB
D 120 x 10^ Btu/hr DMB
O Boiler Data

&
I
~
100
200
'-3
l:

Shading Key to
Coal Types
¦ Utah
9 Illinois
E Wyodak
Q Comanche
	I	
300
Effective HA/SC (10 Btu/hr''ft' )
~1
I
i
o
-------
furnace conditions N0X emissions are higher than would be expected according
to the correlation. This effect may have been caused by the flame stability
problems experienced with the two subbituminous coals under staged combustion
conditions, which were particularly apparent in the cold furnace.
5.6.3 Burner Scale Effects
The aspect of burner scaling was evaluated by testing a 60 x 10® Btu/hr
and a 120 x 10® Btu/hr DMB in a comparable thermal environment. As described
above, this was achieved by adding sufficient insulation to the LWS to
achieve the same furnace exit temperature at 60 x 10® Btu/hr as achieved in
the basic LWS configuration at 120 x 10® Btu/hr. The 60 x 10® Btu/hr DMB,
denoted as the modified configuration, was scaled down from the full-scale
120 x 10® Btu/hr based on maintaining constant velocity. Figure 5-40
summarizes N0X emissions from the full - and half-scale DMBs from three test
coals for a similar thermal environment as determined by furnace exit
temperature. From this figure it can be seen that the larger 120 x 10® Btu/
hr burner tends to produce lower N0X emissions than the comparable 60 x 10®
Btu/hr burner. For operation on Utah coal the differences are not large
(approximately 40 ppm under nominal staged conditions), while for the two
subbituminous coals N0X is substantially higher for the half scale burner.
In general, it might be expected that smaller burners would produce
somewhat higher N0X emissions than a corresponding larger burner. For
constant burner velocities at both.scales, mixing would be expected to occur
more rapidly for the smaller burner dimensions, resulting in correspondingly
shorter flames and higher combustion intensities. However, in practice much
will depend on the details of the burner scaling approach employed and the
corresponding performance of the various devices (e.g. swirlers, impellers)
used to control mixing.
As was indicated previously, the burner zone heat release parameter, HA/
SC varies siightly for the firing conditions of the 60 and 120 x 10® Btu/hr.
The effect of this parameter can be estimated from Figure 5.39 by comparing
the data for the 120 x 10® Btu/hr DMB with the. modified 60 x 10® Btu/hr DMB
5-68
-------
SRt = 1.20
60 x 106 Btu/hr DMB Texit = 1776°F
120 x 106 Btu/hr DMB Texit = 1792°F
,
\o
Wyodak
Comanche
t	r
J	i	I	L
0.5 0.7
0.9
SRr
SRr
SRr
Fi9.ure 5-40. Effect of burner scale on NO emission.
X
-------
fired on Utah coal. If the HA/SC vs. N0X trend established with the short
flame DMBs is assumed to hold, then this figure suggests that the N0X
emission from the 120 x 106 Btu/hr burner is only slightly lower than the
half scale burner, when burner zone heat release is taken into account. This
implies an almost 1:1 scale relationship for the full- and half-scale DMBs.
However, this does not appear to be true for either the Hyodak or Comanche
coals since both produce substantially higher N0X emissions in the reduced
scale burners, and also rank differently in relation to Utah coal. The
reasons for this are not clear, but imply that burner conditions may not
necessarily be optimum for the two subbituminous coals at both burner scales.
5.6.4 Extrapolation to Full Scale
Experimental data for the Circular and DRB burners, obtained both in the
LWS and in field tests on operating boilers, provides an opportunity to
evaluate the scalability of N0X emission performance.
In Figure 5-41 the HA/SC parameter is again used as a means to present
N0X data for the Comanche boiler, and for a corresponding full-scale, circular
burner fired in the LWS. The boiler data presented in Figure 5-41 represents
operation with all burners in service and shows the effect of boiler load,
and the dramatic decrease in N0X resulting from the application of overfire
air. The boiler load data shows a trend with HA/SC which is comparable to
that derived previously for experimental burners and other boilers.
Also shown in Figure 5-41 are data for the 120 x 10® Btu/hr Circular
Burner operating at full and reduced loads with Comanche, Illinois, and Utah
coals. The effect of load with all three coals is consistent, but shows a
much steeper trend with HA/SC. This implies that differences in N0X emission
with load may be due to changes in burner performance in addition to overall
burner zone heat release. If the HA/SC correlation is used to extrapolate
the LWS data to the conditions of the Comanche boiler, then it can be seen
from Figure 5-41 that N0X emissions would be overpredicted by approximately
100 ppm. This suggests either that N0X emissions cannot be correlated in
5-70!
-------
600
en
, i
500 ~
400 -
Q.
Q_
>1
S-
Q
ow 300
o
200
100
	1	
O 120 x 106 Btu/hr
A Reduced Load
~ With Overfire Air
H Boiler Data
/
'	C*'
/	*'
' '	m
/ ~	¦
/ /
d
Open Symbols	- Utah Coal
Half-Open Symbols - Illinois Coal
Closed Symbols - Comanche Coal
		I	

0FA
Open
100	200
Effective HA/SC (103 Btu/hr-ft2)
300
400
Figure 5-41. Correlation of N0x emissions for the Circular Burner.
-------
this way, or that there are differences in burner performance characteristics
at the two scales.
The relationship between DRB N0X emission performance in the LWS furnace
and the Wyodak boiler are presented in Figure 5-42. Here the three lower
data points for the Wyodak boiler represent the effect of boiler load with
one mill out of service. For this data set there is again a consistent
correlation with the HA/SC parameter. Also shown on Figure 5-42 are data
from the 60 x 106 Btu/hr DRB with three coals, and corresponding data for
Pittsburgh #8 coal on an 80 x 1Q& Btu/hr version of the same burner. There
is good agreement between the performance of the two reduced scale burners,
where N0X emissions appear to follow the established HA/SC trend.
The direct extrapolation of the measured DRB performance on Wyodak coal
to boiler conditions is not possible because of the difference in burner
sizes (60 x 106 Btu/hr compared to 120 x 106 Btu/hr in the field). In order
to facilitate this the N0X emission of an equivalent 120 x 106 Btu/hr DRB has
been estimated assuming that the ratio observed for the 60 and 120 x 10® Btu/
hr DMB burners operating on Utah coal is applicable. (The DMB data on Wyodak
coal is not considered sufficiently representative for this purpose). This
yields an estimated N0X level of approximately 250 ppm (compared to 290 ppm
for the half-scale burner), and this point is also shown on Figure 5-42.
Extrapolation of this estimated data point to Wyodak conditions yields
reasonable agreement with the single data point obtained with all burners in
service.
On the basis of the HA/SC correlations, the performance of the 120 x 10®
Btu/hr DMB burner may be extrapolated to the conditions of both the Wyodak
and Comanche boilers for their respective coal types. Figure 5-39 suggests
that this would yield N0X emissions comparable to those already obtained on
these unit's. In this respect it should, however, be noted" that flames
observed in the LWS were confined within the available firing depth of 22
feet, while boiler" observations indicated flame lengths considerably in
excess of this. For the Wyodak unit in particular, flame 1engths were
estimated at 35-44 feet. This Would suggest that burners of the DMB type can
;5-7TJ
-------
500
400
o.
Q.
>1
J-
O
V cm 300
o
Vi
(Jl o
) I
—I	+->
U>	<0
oX 200
100
1 1
O 60 x 106 Btu/hr DRB
1 1
~ 80 x 106 Btu/hr DRB
1
A Estimated 120 x 10^ Btu/hr DRB
1
I


M Boiler Data
~ 1
^ i

1
~ 0
H (All Mills)


I
\
\
\
~ N
\
I

~ •
I One Mill 00S
~
1
* 1
1
Open Symbols - Pittsburgh Coal
1
1
Half-Open Symbols - Utah Coal
1
Closed Symbols - Wyodak Coal
l 1
1 I
0	100	200	300	400
Effective HA/SC (103 Btu/hr-ft2)
Figure 5-42. Correlation of N0x emissions for the DRB.
-------
be applied most successfully to achieve low N0X emissions in confined
situations, or that low N0X performance may be further improved by adjusting
the flames to utilize the available combustion space.
5-74
-------
6.0	S02 REDUCTION POTENTIAL WITH SORBENT INJECTION
The LIMB (limestone Injection Multistage Burner) concept for SO2 control
was evaluated by injecting dry sorbent materials through or around burner
passages. The following parameters were considered:
•	Fuel composition
•	Sorbent composition
•	Injection location
•	Sorbent feed rate
These tests were not intended to	be a comprehensive process optimization;
rather, they were only screening	tests to evaluate possible differences
resulting from burner design or	thermal environment using established
injection locations.
6.1	Injection Configurations
Sorbent injection tests for SO2 control were completed for the following
furnace configurations:
9 limited 60 x 10^ Btu/hr DMB with coal impeller in both baseline and
insulated LWS
§ 120 x 10® Btu/hr Circular burner
•	120 x 106 Btu/hr DMB
Each burner was operated at nominal full load with 20 percent excess air.
A total of six different injection locations were considered. The two
near burner locations, with the coal and through nozzles located on the axis
of each tertiary port for the DMBs, are illustrated in Figure 6-1. The four
locations in the furnace above the burner are shown in Figure 6-2. Table 6-1
lists the locations used for each burner configuration.
-------
Sorbent
Twin
Screw
Feeder
Scale
valve
yen curl
Compressor
Flow Meter
Metered
Sorbent
to {b) or (c)
s) Sorbent Feed System
Tertiary Ports
Concentric
Sorbent Infection
Pipes
Sorbent from
Feeder
4-way Splitter
LWS Mounting Plate
Burner Exit
b) Sorbent Injection through
Tertiary Air Parts
Coal Feeder
EL
Metered
/ >-
To
Burner
Hot-Air
Pulverizer
Sorbent frcfi
Feeder
Exhauster
c) Sorbent Injection with Coal
Figure 6-1.
Schematic representation of sorbent injection system.
6-2
-------
• - - o —o -
— O o o o
231
19'
¦O O
¦o o
4-
23' -
19" "
8"
4>
Burner
~
\
Furnace Locations
Figure 6-2. In-rfurnace "sorbent injection locations.
6-3
-------
TABLE 6-1. SORBENT INJECTION LOCATIONS EVALUATED FOR
EACH BURNER CONFIGURATION
Sorbent Injection
Location
60 x 106 Bt
u/hr DMB
120xl06 Btu/hr
Ci rcular
120xl06 Btu/hr
DMB
Baseline LWS
Insulated
With Coal
X
X
X
X
Through Tertiary
Air Ports
X
X

X
4' above burner
X
X


8' above burner
X
X
X
X
19' above burner


X
X
23' above burner
(LWS nose)


X






-------
6.2	Test Results
6.2.1 60 x IP6 Btu/hr DMB
A series of tests were conducted to evaluate the potential of SO2
reduction by injecting dry sorbents. Two processed calcium based sorbents,
Vicron 45-3 1 imestone and Col ton hydrated lime were used. Four different
injection locations were evaluated with the 60 x 10® Btu/hr DMB.
1.	With the coal at the outlet of mill exhauster
2.	Through nozzles located on the axis of each tertiary air port
3.	Lower level overfire air ports - 4 feet above burner center!ine
4.	Middle level overfire air ports - 8 feet above burner centerline
Extensive sorbent injection tests were conducted with Illinois Coal.
Illinois coal, is a high sulfur coal {3.76 percent S dry) and has been used
in previous LIMB development work at EER. Figure 6-3 summarizes the effects
of sorbent injection on SO2 capture with Illinois Coal. Figure 6-3(a) shows i-
injection of Vicron and lime through the tertiary ports. Injection of
hydrated lime yielded the highest capture of 53 percent compared to 48.5
percent with Vicron at a calcium-to-sul fur molar ratio of 2.0. Injection of
Vicron with the coal yielded slightly lower capture, 44.5 percent at Ca/
S = 2.0. Figure 6-3(b) summarizes the results of sorbent injection through
two different overfire- air (0FA) locations. The level of 0FA ports had no
significant effect on sulfur capture by the injection of hydrated lime.
Injection of Vicron through the lower ports yielded higher capture than when
injected through the upper ports.
An abbreviated series of tests were conducted for Utah, Wyodak, and
Comanche coals with sorbent injection, summarized in Figure 6-4. The highest
capture was achieved with injection of hydrated lime through the lower 0FA
ports for each coal. Injection with this configuration yielded SO2 reduction
of 50 and 48 percent at Ca/S = 2.0 for Utah and Comanche, respectively.
Sulfur capture was significantly less for the Wyodak coal with only
6-5
-------
Noalnil Conditions:
Fuel - Illinois
Nominal Conditions:
Fuel - Illinois
Firing Rate ¦ 58.5 * 1.1 x 10 Btu/hr Register Post ion (° Open)
srb • o.6a 1 o.ou
SHt ¦ 1.2) * 0.021

SORBENT
INJECTION LOCATION
o
Vlcron
Tertiary Ports
n
Vlcron
Utth coal
A
Hydratejl^
Tertiary Ports
2	3	4
! Ca/S Molar Ratio j
a) Injection through burner passages
Inner - 20 CCV
Outer - 10° CU
Firing Rate - 59.0 . 1.5 x 10 Btu/hr
SRj, - 0.67 t 0.01
SRr
1.20 t 0.02
f

Sorbent
Location
O
Vlcron
OFA - Lower
~
Line
OFA - Lower
¦
Line
OFA - Middle
•
Vlcron
OFA - Middle
a
2	1
j Ca/S Molar Ratio |
b) injection tiirouyh Of A ports.
Figure 6-3. Summary of sorbent injection with 60 x 10 Btu/hr DMB
Illinois coal.
-------
cn
i
/
/
f
7
Nominal Conditions:
Firing Rate = 58.0 ! 0.7 x 10^ Btu/hr
SRP = 0.69 + 0.01	SRt = 1.20 t 0.015

LOCATION
SORBENT
FUEL
A
~
Terti aries
0FA (Lower)
Vicron
Lime
Utah
~
¦
Tertiaries
0FA (lower)
Vicron
Lime
Wyodak
0
Tertiaries
OFA (lower)
Vicron
Lime
Comanche
Ca/S Molar Ratio
Figure 6-4. Summary of sorbent injection with 60 x 10° Btu/hr DMB burner.
-------
41 percent at a comparable stoichiometry. Injection of Vicron limestone
through the tertiary air ports reduced SO2 by 42 percent with Utah Coal.
Comanche and Wyodak had significantly lower SO2 reductions with 35 and 32
percent respectively.
A series of tests were initiated to evaluate the effect of thermal
environment on SO2 reduction potential of the short flame 60 x 10® Btu/hr DMB
with sorbent injection. The initial tests with Illinois coal are summarized
in Figure 6-5 for sorbent injection through the tertiary and lower overfire
air ports but with furnace temperatures increased through installation of
additional insulation. For these injection locations the SO2 capture was
significantly less in the insulated LWS compared to results in the baseline
furnace. Capture was higher for injection of hydrated lime through the lower
overfire air ports (OFA), 37 percent at Ca/S = 2.0, than the other
combinations. This is almost 7 percent less than achieved with the same
conditions in the baseline furnace. Injection of the Vicron limestone
through the tertiary air ports and the lower OFA ports produced similar SO2
reduction, 26 percent at Ca/S = 2.0. This is much less than the 41.5 percent
achieved in the baseline furnace for the same conditions. These results
indicate the sensitivity of the SO2 capture process by sorbent injection to
thermal environment. Additional SO2 capture data, for sorbent injection
through the middle OFA ports (8 feet above the burner center), are presented
in Figure 6-6. Here the SO2 capture data for the insulated furnace are
compared directly with the previous data for baseline conditions.
For the insulated furnace, the highest capture was obtained with
injection of hydrated lime through the middle overfire air ports, yielding
53.3 percent capture at Ca/S = 1.93. Slightly 1ower SO2 reduction, 50.3
percent, was obtained with Vicron. Vicron injected with the coal resulted in
lower SO2 capture, 45 percent at Ca/S = 2.0. The thermal environment had the
greatest effect on the injection of Vicron through the middle OFA ports with
capture increasing by about 12.5 percent at Ca/S = 2.0 in the insulated
configuration. For hydrated lime injected through the middle OFA ports, the
overall trend indicates generally no difference due to thermal environment.
6-8
-------
SORBENT
O ¦ VICRON
Q HYDRATED LIME
¦ HYDRATED LIME
• VICRON
LOCATION
TERTIARY PORTS
TERTIARY PORTS
LOWER OFA
LOWER OFA
NOMINAL CONDITIONS:
FUEL - ILLINOIS	|
FIRING RATE = 59.6 t
i .2 x :o°
Btu/hr.
0.69 + 0.017
+ 0.015
SRb
SRt
1.21

CL
o
CM
'O
to
Ca/S' Molar Ratio
Figure 6-5. S0? reduction potential with sorbent injection
of the 60 x 10® Btu/hr DMB in the insulated LWS.
-------
NOMINAL CONDITIONS
FUEL - ILLINOIS
FIRING RATE =59.0
SRr * 0.68 t 0-019
+
1,5 x 135 Btu/hr
SR
T
1.21 t 0.018
Q VICRON
O HYDRATED LIME
O VICRON
OFA - MIDDLE
OFA - MIDDLE
MILL (WITH COAL)
60
50
** . 40

Q.
fd
O
CM
O
VO
/
NOTE: Shaded Symbols
Indicate Insulated
LWS
Ca/S Molar Ratio
Figure 6-6. Summary of S0? reduction potential with
60 x 10 Btu/nr DMB burner with Illinois
coal.
6-10-
-------
This same trend is found for the injection of Vicron with the coal with no
change in capture over the tested range.
The effects of sorbent injection in the insulated furnace were also
evaluated for two alternate fuels Utah and Wyodak. Figure 6-7 summarizes the
effects of sorbent injection for Utah coal. The highest SO2 capture 47.5
percent was obtained with the injection of lime through the OFA port at Ca/
S = 2,0 for both temperature/injection location combinations. Substantially
lower capture was obtained with injection of Vicron through the tertiary
ports, with a 40 percent SO2 reduction at Ca/S = 2.0 in the baseline furnace.
Increasi ng the furnace temperature with additional insulation significantly
decreased the SO2 reduction to 29 percent at Ca/S = 2.0. Figure 6-8
summarizes the effects of sorbent injection on SO2 reduction with Wyodak host
coal through the OFA ports. Slightly higher capture was obtained in the
insulated furnace, 48 percent compared to 41 percent at Ca/S = 2.0 with
injection of hydrated lime through the OFA ports. The difference in capture
was even less {less than 4 percent change) for other calcium-to-sulfur molar
ratios.
6.2.2 120 x 106 Btu/hr Circular Burner
A comprehensive series of tests were conducted with the circular burner
to evaluate the SO2 reduction potential through injection of two dry
sorbents, Vicron limestone and hydrated lime. Four different injection
locations were utilized corresponding to OFA ports 8 feet, 19 feet, and 23
feet above the burner center!ine, and with the coal. These locations
correspond to a temperature range of about 2200-2300°F at the 8 foot level to.
approximately 1800-1850°F at the 23 foot level.
A brief series of tests were completed with Utah coal and is summarized
in Figure 6-9. Injection of hydrated lime through the 8 foot ports resulted
in SO2 capture of about 35 percent at Ca/S = 2.0.
Injection of Vicron limestone with the coal resulted in significantly
lower SO2 reduction, about 10 percent at Ca/S = 2.0. This extremely low
6-11
-------
NOMINAL CONDITIONS:
FUEL - UTAH
FIRING RATE=58.2tl.4x1Q6Btu/hr
SRr = 0.69+0.008
SRj = 1.21 to.02

CV
u.
3 .
+->
£X
o .
Ul
o
A
S0RBEMT
VICR0N
HYDRATED
RYDRATED
IME	
VICR0N
LOCATION
TERTS
UNINSULATED
LOWER OFA
UNINSULATED
MIDDLE OFA - INSULATED
TERTS
INSULATED
/
/
/
Ca/S Molar Ratio i
Figure 6-7.
Summary of sorbent injection with 60 x 10 Btu/hr
burner with Utah coal.
6-12
-------
SORBENT - HYDRATED LIME !
LOCATION	|
o OFA - LOWER '
• OFA - MIDDLE .. i
NOMINAL CONDITIONS;
FUEL - WYODAK ^	,
FIRING RATE = 57.4 t 1-0x10®
Btu/hr
SRg = 0.68 t 0.019
SRj = 1.20 t 0-02
70
60
Hot" LWS
50
40
Coltf". LWS.?
20
10
0
0
Ca/S Molar Ratio
' }
[Figure 6-8. Effect of sorbent injection on S02 reduction
j	potential with Wyodak coal.
6-13
-------
Fusl » Utah
Firinc Rate = 119.4 ± 2.5 x 105Btu/hr
SRj = 1.20 ! 0,1

SORBENT
LOCATION
0
Q
Vicron Limestone
Hydrated Lime
Mill
3'

r
/
/
/
Figure 6-9.
3	 4
Summary of sorbent injection with
J20 x 10° Btu/hr circular burner
wftfi Ofali coal. ~				
6-1 A-
-------
capture is probably due to the intense temperature (>2300°F) to which the
sorbent is exposed in the center of the flame zone.
A more extensive series of tests were conducted firing the high sulfur
Illinois coal with the Circular burner. These results are summarized in
Figure 6-10. Injection of Vicron limestone yielded similar SO2 reduction,
regardless of location. The best capture was achieved at the 19 foot and
23 foot level with capture rates of about 38 percent at Ca/S = 2.0. This is
only slightly higher than the capture obtained when injected with the coal or
at the 8 foot level which resulted in about 33.5 percent reduction.
Injection of hydrated Time yielded increased capture rates of 42 percent at
both the 8 foot and 19 foot level at Ca/S = 2.0. However, at the 23 foot
level , capture was significantly lower with only 32 percent at Ca/S = 2.0.
These results indicate a trade-off between residence time.in the optimum
sulfation window and the maximum temperatures to which the sorbent is
exposed,
6.2.3 120 x 106 Btu/hr DMB
A series of tests were also conducted to evaluate the potential of SO2
reduction with the 120 x 106 Btu/hr DMB. Three processed calcium-based
sorbents were utilized, Vicron 45-3 limestone, Col ton hydrated lime, and
Genstar Type-S pressure hydrated dolomitic lime (PHDL). Four different
injection locations were evaluated:
1.	With the coal at the outlet of the mill exhauster.
2.	Through nozzles located at the axis of each tertiary port.
3.	Lower level overfire airports - 8 feet above burner center!ine.
4.	Upper level overfire airports - 19 feet above burner center!ine.
An abbreviated series of tests were conducted with Utah coal and is
summarized in Figure 6-11. Two injection locations, the tertiary air ports
and the 8 foot level were evaluated. Figure 6-11 (a) represents injection
through the tertiary air ports. There is no significant difference between
6-15
-------
Fuel - Illinois
Firing Rate = 120.5 - 1.04xl0"Btu/hr
SRj = 1.19 ± 0.01
Injection
Location

with coal
8 ft.
19 ft.
23 ft.
Sorbent
A
o
(ft
•
Vicron Limestone

~
n
¦
Hydrated Lime
X
QJ 30

/
1	2	3
Ca7S Molar Ratio
a) Near Burner Injection
i	
Ca/S Molar Ratio
b) Upper Furnace Injection
5-
Figure 6-10. Summary of sorbent injection with 120 x 10 Btu/hr circular
burner with Illinois coal.
-------
cr» ,
I
O VICRON LIMESTONE
~ HYDRATED LIME
O PHDL

70
60
50
40
a>
¦i-
3
;->o-
'o 30
' CSJ
' o
oo
20
10
NOMINAL CONDITIONS:
FUEL = UTAH
FIRING RATE = 119.5
SRB = 0.71 t 0.02
SRt = 1.20 I 0.01
1.3 x 10 Btu/hr
0 1	2	3 ,4
* Ca/S Mol ar Ratio •/
a a) Tertiary Port Injection
« 30
Ca/S Molar Ratio
b) 8 Foot Level
Figure 6-11. Summary of sorbent injection with the 120 x 10® Btu/hr DMB fir-
ing Utah coal.
-------
the Vicron limestone or the hydrated lime with SO2 capture at about
23 percent at Ca/S = 2.0. Figure 6-11(b} summarizes injection at the 8 foot
level. There is no significant change in injection of Vicron limestone at
this level compared to the near burner tertiary air locations. The 8 foot
level yielded slightly higher capture with hydrated lime, up to about
27.5 percent at Ca/S = 2.0. The highest capture was obtained with injection
of the PHDL at the 8 foot level with capture of about 40 percent at Ca/
S = 2.0.
Extensive sorbent injection tests were conducted with the Illinois coal,
and results are summarized in Figure 6-12. As with the Utah coal, there is
no difference between injection of Vicron limestone or hydrated lime through
the tertiary air ports although overall capture is slightly higher, 27
percent at Ca/S = 2.0. Injection of Vicron limestone with the coal is
slightly lower with 25 percent SO2 reduction at Ca/S = 2.0. Figure 6-12(b)
shows the improved results obtained with injection into the upper furnace.
The capture rates are higher in each case than with the Utah coal. Vicron
limestone injected at the 8 foot level achieved slightly higher capture than
either "near burner" location, 29 percent at Ca/S = 2.0. Injection with
hydrated lime resulted in a more significant increase with SO2 capture of
about 33 percent at the 8 foot level and 40 percent at the 19 foot level at
Ca/S = 2.0. Once again, the highest SO2 reduction is achieved with the PHDL
with capture of 40 percent at Ca/S = 1.0. Higher Ca/S molar ratios with the
PHDL would result in increased capture, but due to the sorbent properties and
higher feed rates required, it was not possible to inject at any higher Ca/S
ratios at full load operation.
A brief series of tests were conducted to determine the effects of
thermal environment on SO2 reduction potential by injecting sorbent at a
reduced firing rate of 65 x 106 Btu/hr with Illinois coal. These results are
summarized in Figure 6-13. For each sorbent/injection location combination,
SO2 capture improved significantly. With injection of hydrated lime into the
8 foot level the capture increased from 40 at full load to 50 percent at the
reduced load. Injection of Vicron limestone into the 8 foot level also
increased capture rate from 29 to 40 percent. Hydrated lime injected through
6-18
-------
NOMINAL CONDITIONS:
FUEL - ILLINOIS	,
FIRING RATE = 120.5 t 1.4 x 106 Btu/hr
SRb = 0.70 ! 0.01
SRt.= 1.19 t 0.01

S0R6INT
LOCATION
o
VICR0N LIMESTONE
TERTIARIES
~
HYDRATED LIME
TERTIARIES
•
VICR0N LIMESTONE
with coal

S0RBENT
LOCATION
o
VICR0N LIMESTONE
8 ft
A
PHDL
19 ft
O
HYDRATED LIME
8 ft
¦
HYDRATED LIME
19 ft

0)
u
70
60
50
40
a.
-------
NOMINAL CONDITIONS:
FUEL - ILLINOIS	fi
FIRING RATE = 65.7 t 1.3 x 10 Btu/hr
SRg = 0.70 1 0.01
SRj = 1.19 t 0.01

SORBENT
LOCATION
o
VICRON LIMESTONE
8 ft.
~
HYDRATED LIME
**-
CO
A
PHDL
8 ft.
¦
HYDRATED LIME
Tert. Ports
•
VICRON LIMESTONE
with coal
Ca/S Molar Ratio
Figure 6-13,
Summary of sorbent injection with
staged 120 x 105 Btu/hr. DMB at
reduced fifing rate.
6-20
-------
the tertiary ports Increased capture from 27 to 40 percent as a result of the
decreased firing rate. SO2 capture increased to 36 percent with injection of
Vicron limestone with the coal. The highest SO2 reduction was again obtained
with the PHDL with 62.5 percent reduction at Ca/S = 2.0 for injection at the
8 foot level.
6.3	Discussion of SO2 Removal Performance
Previous sections have presented detailed SO2 removal results for the
experimental burners investigated in this program. A wide range of SO2
removal rates have been encountered in the study, depending upon the
parti cul ar experi mental conditions of i ndi vi dual tests . A summary of
relevant S02 removal data is presented in Table 6-2 for limestone and
hydrated lime injection at Ca/S = 2, and operation on Illinois coal,
parameters for which the largest data base was obtained. This summary table
al so covers the effects of burner design, firing rate, furnace insulation,
and sorbent injection 1ocation. In order to elucidate those parameters which
have the greatest impact on the sorbent injection process some further
discussion is necessary.
Over the past several years considerable work has been conducted at
laboratory, bench, and pilot scales, in the study of both fundamental and
practical aspects of sorbent injection. While the process is not completely
understood, factors which are known to significantly affect SO2 removal
performance include:
•	Sorbent type
•	Temperature at the injection location
•	Furnace temperature profile (quench rate)
•	Coal type
•	Mixing
The following sections will draw on the previous data presentation to
illustrate the impacts of the above parameters.
6-21
-------
TABLE 6-2. SUMMARY OF PERCENTAGE S02 REMOVAL DATA-ILLINOIS COAL,
INJECTED Ca/S MOLAR RATIO = 2.

60 x 1G6 Btu/hr DMB
120x10®
Btu/hr
Circular
120xl05
Btu/hr
DMB
oOFDent
Injection
Location -
Baseline LWS
Insulated
CaC03
Ca(OH) 2
CaC03
Ca(0H)2
CaC03
Ca{0H>2
CaC03
Ca{0H)2
With Coal
44

44

35

27

Through
Tertiary
Air Ports
48
53
25
30
-

27
27-
4' Above
Burner
42
51
25
38




8' Above
Burner
37
47
50
53
33
42
28
33
19' Above
Burner




38
43

38
23' Above
Burner
(LWS' nose)




38
32


6-22
-------
6.3.1 Sorbent Type
The data presented in Table 6-2 indicate that, for comparable firing and
injection conditions, higher levels of SOg removal are generally achieved
with hydrated lime than with limestone. Other limited data obtained with the
pressure hydrated dolomitic sorbent (e.g. Figure 6-12) indicate that this
material significantly outperforms both of the high calcium sorbents. This
general ranking of sorbent materials is consistent with that found in other
studies, and Figure 6-14 presents data for the test sorbents obtained in a
bench scale furnace facility under carefully controlled experimental
conditions. These data suggest that the Col ton hydrate is only marginally
more effective than Vicron limestone. It will be noted, however, that the
majority of the SO2 capture data obtained in the LWS furnace is considerably
in excess of that reported in Figure 6-14 for more idealized conditions. The
reasons for this are believed to relate to other parameters such as injection
temperature, furnace thermal profile, and mixing.
6.3.2 Injection Temperature
In the execution of the experimental program the effect of sorbent
injection temperature on SO2 removal was evaluated by two means;
•	The use of injection locations close to the burner and at different
elevations above the burner.
•	Increasing mean furnace temperatures through additional insulation.
For the 60 x 10® Btu/hr 0MB the results of Table 6-2 and the preceding
discussion indicates that:
•	For equivalent injection locations increasing furnace temperature
decreases SO2 capture.
•	Optimum S0£ capture is achieved by injection at hi gher furnace
elevations as furnace temperature is increased.
6-23
-------
O Vicron
O Colton Hydrate
O Genstar Dolomitic Hydrate(s)
, 20 -
2250 F Injection
1000 F/sec Quench
3
2
1

Ca/S Molar Ratio
Figure 6-14. Bench scale data comparing $0,, removal performance
for different sorbents.
6-24
-------
These results imply that there is an optimum injection location, which is
determined by temperature for each furnace/burner configuration. The 60 x
106 Btu/hr DMB data from Table 6-2 are plotted in Figure 6-15 as a function
of temperature at the sorbent injection plane. Temperature values assigned
to the different injection locations and levels of furnace insulation are
estimated from limited available measurement data, and from simple heat
transfer models of furnace performance, and are intended to represent the
maximum temperature experienced by the sorbent. Figure 6-15 shows a
reasonable correlation between SO2 capture and injection temperature, and
indicates an optimum temperature of approximately 2200°F for both limestone
and hydrated lime. For the baseline (cold) LWS this location corresponds
approximately to the tertiary port location, while for the insulated furnace
this temperature occurs at approximately 8 feet above the burner. This kind
of dependency on injection temperature is again expected from smaller scale
sorbent injection studies.
It will be noted, however, that the Table 6-2 data for the two 120 x 10®
Btu/hr burners do not readily fit into the plot on Figure 6-15, either in
terms of maximum SO2 removal, or in terms of the relative location at which
maximum capture occurs. This is in spite of the use of additional insulation
in one 60 x 106 Btu/hr case in an attempt to match thermal conditions with
the higher firing rate. The reason for this behavior is believed to be due
both to the detailed temperature profile in the furnace and to sorbent/flue
gas mixing.
6.3.3 Temperature Profile
A further parameter which has been shown to significantly impact SO2
removal performance is that of the temperature profile in the furnace. This
is often expressed as a mean quench rate or the rate of temperature decay
over the internal 2200-1600°F. Results of studies in small scale furnaces
are presented in Figure 6-16 to illustrate the effects of quench rate and
injection temperature on SO2 removal, for limestone and hydrated lime
sorbents. The figure indicates that as conditions move from isothermal to
• .6-25/
-------
cn
TO
o>
60

40
a;
u
3
-P
a_
ro
O
CVJ
o
00
20
O GaC03
A Ca(0H)2
Solid Symbols—Insulated Furnace
Open Symbols—Baseline Furnace
	I	I	
_L
i.
Figure 6-15.
1800	2000	2200	2400
Estimated Temperature at Injection Plane (°F)
Estimated relationship between injection temperature and
S09 capture—60 x 10® Btu/hr DMB, Illinois coal.
-------
1 1 1 1
_ Quench Rate
1 1 1 1


	400°F/sec




/ ^ \ \
• / \

/ \
/ / \ X- X
i
/
/
1/
1
1
\
\
\
\
	L



CaC03
Ca(0H)2
I I 1 1
I I 1 i
1800 2000 2200 2400	1800 2000 2200 2400
Injection Temperature (°F)
Figure 6-16. Small-scale data showing the effect of thermal
conditions ion/JSO^ capture;'
-------
rapid quench the maximum SO2 removal is significantly reduced, and optimum
temperatures tend towards higher values.
In the context of the LWS furnace the applicable quench rate is
estimated to be very low (below 200°F/sec5, such that long mean residence
times are available in the optimum temperature range. This is particularly
true for the cold furnace condition and low firing rates. In comparing the
120 x 10® Btu/hr data with 60 x 10® Btu/hr data in the insulated furnace, it
is clear that although mean temperature levels are approximately matched, the
effective quench rate will approximately double at the higher firing rate.
This is believed to be the primary reason for the decrease in optimum SO2
removal at 120 x 10® compared to- 60 x 10® Btu/hr. This point can be further
illustrated by data obtained with the 120 x 10® Btu/hr burner at half load.
In this mode of operation the furnace thermal conditions are comparable to
those at 60 x 10® Btu/hr in the baseline LWS. Figure 6-13 shows that for the
120 x 10® Btu/hr DMB operated at half load, and Ca(OH}2 injection at the 8
foot elevation, SO2 capture at Ca/S = 2 is increased from 33 percent to 50
percent. This value is comparable to the 47 percent achieved with the 60 x
10® Btu/hr DMB in the baseline LWS with equivalent injection conditions.
6.3.4 Mixing
Mixing of the injected sorbent stream with the flue gases has been shown
to be an important parameter in ensuring effective SO2 removal. Although
specific mixing studies were not conducted in connection with this program,
mixing is believed to be important in influencing some of the results. The
actual mixing of the sorbent jets with the bulk of the flue gases is not
considered to be a problem, since injectors were designed for adequate
penetration and coverage in the relatively low velocity flue gas stream.
However, the general flow field in the LWS is characterized by large
recirculation zones both above the burner and in the ash hopper as a result
of the comparatively unconfined burner environment. For small burners (60 x
10® Btu/hr) and long flame configurations, the upper recirculation zone may
be particularly strong and may re-entrain a large fraction of the flue gas
flow. For the moderate thermal environment which characterizes many of the
6-28,:
-------
LWS configurations, this can result in relatively long residence times, and
even recycling of a significant fraction of the sorbent material, through
relatively optimum thermal conditions. This effect is believed to be
responsible for the high optimum SO2 removal rates which are achieved,
compared to small-scale data in more idealized (Figure 6-14) and isothermal
(Figure 6-16) conditions.
6.3.5	Coal Type
The impact of coal type on SO2 capture is believed to be due primarily
to coal sulfur content and the resulting SO2 concentration. However, small
scale data indicate that for SO2 concentrations above approximately 1000 ppm
the impact on SO2 removal is relatively small. Indeed, in the preceding
presentation of LWS test data only comparatively small differences were
apparent for the limited data on different coal types. In general, slightly
higher SO2 removal rates were observed for the higher sulfur Illinois coal,
compared to the lower sulfur Utah, Comanche, and Wyodak fuels which tended to
yield comparable results. Some of these results may also have been
influenced by small changes in thermal conditions, e.g. as a result of firing
bituminous vs. subbi tumi nous coals. However, it should be noted that the
majority of the coal comparison data was obtained at 60 x 10® Btu/hr, where
$02 capture values were relatively high and where capture may have been
dominated by combinations of the thermal and mixing criteria referenced
above. This may have masked some of the differences between the different
coal types which would be expected to be more apparent. The limited data
available for 120 x 10® Btu/hr operation suggests that the relative
differences between coal types (e.g. Illinois vs. Utah in Figures 6-10 and
6-11) might become more pronounced under different thermal and flow
conditions where the overall level of SO2 removal is reduced.
6.3.6	Extrapolation to Full Scale
The preceding discussion has indicated that the SO2 removal process
depends upon a complex interaction between a large number of physical,
thermal, and chemical parameters. For this reason it considered not to be
6-29
-------
possible to directly extrapolate the data obtained in the LWS to full-scale
boiler systems. At full scale the complex multiple burner interactions and
different furnace geometry and confinement will result in flow fields and
thermal profiles substantially different from those encountered in the LWS.
Consequently, 1t is believed that each boiler application must be
characterized on an individual basis, and sorbent injection locations and
injection devices selected to provi de thermal condi ti ons and mixing
characteristics which are optimum for that application.
6-30
-------
7.0	REFERENCES
1.	Chen, S. L,» M. P. Heap, D. W. Pershing, and G. B, Martin. The Fate of
Coal Nitrogen During Combustion, Fuel, 61, 1218 (1982).
2.	Crawford, A. R., Manny, E. H., and Bartok, W. Field Testing: Application
of Combustion Modifications to Control N0^ Emissions from Utility
Boilers, EPA-650/2-74-066, NTIS PB 237 344, June 1974.
3.	Zallen, D. M., Gershman, R., Heap, M. P., and Nurick, W. H. The
Generalization of Low Emission Coal Burner Technology. _In: Proceedings
of the Third Stationary Source Combustion Symposium; Volume II Advanced
Processes and Special Topics, EPA-600/7-79-Q50b, NTIS PB29254G,
February 1979.
4.	Leikert, K. and Michelfelder, S. Operating Experience and Field Data of a
700MW Coal-Fired Utility Boiler with Retrofit Low NO Staged Mixing
Burners. In: Proceedings of the Joint Symposium of Stationary
Combustion NO Control, Volume II, Utility Boiler NO Control by
x	x
Combustion Modification, EPA-600/9-81-028b, NTIS PB81-236127, July 1981.
5.	Beer, J, M. and Chigier, N. A.. Combustion Aerodynamics. Applied
Science Publishers Ltd., 1972, pp. 109-112.
6.	Gupta, A. K,, Lilley, D. G., and Syred, N. Swirl Flows. Turnbridge
Wells: Abacus Press, 1984, pp. 1-4.
7.	Pohl, J. H., Chen, S. L., Heap, M. P., and Pershing, D. W, Correlation
of NO Emissions with Basic Physical and Chemical Characteristics of
x
Coal, In: Proceedings of the 1982 Joint Symposium on Stationary
Combustion NO Control, Vol. II, EPA-600/9-85-022a, NTIS PB85-235612,
x
July 1985.
7-1
-------
—-- i TECHNICAL REPORT DATA
. - - (Please read Instructions an the reverse before completing)
1, REPORT NO. J 2. " '
EPA-600/7-89-015a
3, RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
Field Evaluation of Low-Emission Coal Burner Tech-
nology on Utility Boilers; Volume I. Distributed Mix-
ing Burner Evaluation
5. REPORT DATE
December 1989
6. PERFORMING ORGANIZATION CODE
7, AUTHOR(S)
' A. R. Abele, G. S. Kindt,and R. Payne '(EERC); and
P. W. Waanders (Bab cock and Wilcox)*
a. PERFORMING ORGANIZATION report no.
9, PERFORMING ORGANIZATION NAME AND ADDRESS
Energy and Environmental Research Corporation
18 Mason
Irvine, California 92718
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3130*
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711
13. TYPE of REPORT AND PERIOD COVERED
Final; 9/78-6/86
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes _/\EERL project officer is P. Jeff Chappell, Mail Drop 63, 919/541"
3738.\(*) Contract with Babcock and Wilcox, P, O. Box 351, Barberton, OH 44203-
n3R1. X*
is. ABSTRACTaiTlie rep0rj- gives results of a study in which NOx emissions and general
combustion performance characteristics of four burners were evaluated under ex-
perimental furnace conditions. Of primary interest was the performance of a low-
NOx Distributed Mixing burner (DMB), which was tested in a nominal full-scale (120
million Btu/hr or 35MW) version and in a corresponding half-scale version. Perfor-
mance was compared against a half-scale commercial low-NOx Dual Register burner
(DRB) and a 120 million Btu/hr commercial Circular burner. The report documents
the performance of each burner type over a wide range of firing conditions and for
different bituminous and aibbituminous coals. Additional test program goals were to
provide information relating to the effects of burner design, burner scale, and ther-
mal environment on NOx emission performance. Full- and half-scale DMB perfor-
mance was compared under equivalent thermal conditions; the DMB was tested under
two levels of furnace insulation; results from the DRB and Circular burner were
compared to field data from two utility boilers operating with corresponding burner
designs and coal types. A burner zone heat liberation rate parameter was used to
compare the relative performance of the different burners under the various firing
conditions.
M. KEY WORDS AND DOCUMENT ANALYSIS
a. . DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Pollution Calcium Oxides
Coal Sorbents
Combustion Dolomite
Nitrogen Oxides Electric Utilities
Sulfur Dioxide
Calcium Carbonates
Pollution Control
Stationary Sources
Low-NOx Burners
13B
2 ID 11G
21B 08G
07B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)'
Unclassified
21. NO. OF PAGES	
riS7— j
20. SECURITY CLASS (This page}
Unclassified
22. PRICE
EPA Form 2220-1 {9-73)
-------