Bps-06 NOx Control

BURNING PRB COAL
Minimize NOx using

only combustion control
You may not have to spend big bucks on an SCR system to reduce your NOx
emissions. The retrot of a 600-MW opposed-wall-red utility boiler with
low-NOx, dual air zone burners and overre air cut the Wyoming PRB
coal burners NOx output nearly in half. The key to the projects success,
from the design stage through nal testing and boiler tuning, was CFD
modeling.
By Craig A. Penterson and Kenneth R. Hules, ScD, Riley Power Inc., a Babcock Power Inc. Company
acing tighter fleetwide NOx limits, a
midwestern utility found itself having
to decide which emissions-control
technologies to install at which plants. As
part of the process, the utility did a technical
and economic analysis of NOx-reduction
options. The analysis suggested that upgrading the combustion control of one of its
plants would avoid having to add an expensive selective catalytic reduction (SCR) or
selective noncatalytic reduction(SNCR) system at the plant and be sufficient to meet
eetwide NOx-reduction goalsbut only if
doing so reduced the plants NOx output by
half.
Babcock Power Inc., through its Worcester, Mass.based subsidiary Riley Power Inc.
(RPI), took on the challenge of reducing the
plants NOx emissions from 0.30 lb/million
Btu (mmBtu) to 0.155 lb/mmBtu, while
maintaining CO levels below 100 ppm and
carbon-in-ash levels below 0.5%.
The boiler in question is opposed-wallfired, originally burned subbituminous
Powder River Basin (PRB) coal exclusively, and came equipped with 56 low-NOx,
dual-register, first-generation Babcock &
Wilcox burners at four elevations. However, due to high localized furnace heat
release rates and excessive slagging, 16 of
the burners (4 per elevation) were removed
from service following startup. The boiler
was designed to generate 4,550,000 lb/hr
of main steam at 2,650 psig and 1,005F,
and 4,281,000 lb/hr of reheat steam at 602
psig and 1,005F. At full load, the plants
turbine-generator produces approximately
600 MW.
A clean-burning design
Major modications to the boilers existing
combustion hardware included installing
upgraded burners, new overre air (OFA),
and rearranging the burners to allow those
at the upper elevation to take their combustion air from the lower three windbox compartments. The OFA nozzles were installed
in the uppermost windbox compartment.
The burner selected for the project uses

RPIs latest low-NOx, dual air zone, controlled-combustion venturi (CCV) burner
technology (Figure 1). The venturi coal
nozzle, low-swirl coal spreader, and ame
stabilizer ring produce a well-attached,
fuel-rich amethe fundamental necessity
for minimizing the formation of both fuel
and thermal NOx.
The use of OFA further stages the combustion process and produces additional
NO x reduction. Typically, up to 25% of
the combustion air can be diverted from
the main burner and introduced above the
top elevation of burners. This reduces the
stoichiometry at the main burner zone and
creates a reducing environment for combustion. Air staging is limited (or not used
at all) on boilers burning high-sulfur coals
and on boilers designed for supercritical
operation due to the possibility of lower
furnace waterwall corrosion.
Separate on/off dampers within each
OFA section control air ow through each
1. The low-NOx, dual air zone burners schematic (L) and hardware (R)
Shroud actuator
(automatic)
Flame scanner
Venturi coal nozzle

Tertiary air swirl
vanes (adjustable)
Ignitor
Burner head
Flame stabilizer ring
Secondary air diverter

Tertiary air diverter
Low-swirl coal spreader
Secondary air flow

control damper
Burner air shroud

Secondary air swirl
vanes (fixed)
Source: Riley Power Inc.

48
www.powermag.platts.com
POWER
| October 2005
BURNING PRB COAL

of the compartments. This results in better
control of the penetration and mixing of the
overfire air over a range of operating
loadsa key factor in optimizing combustion and minimizing CO production
throughout the load range.
The NOx-reducing features of the burner
design include dual combustion-air zones,
the low-swirl coal spreader, and the ame
stabilizer ring. The zoning divides the combustion air into secondary and tertiary air
their design, to obtain data for the furnace

model, and to obtain the burner settings
needed to produce the desired low-NO x
flow patterns for initial settings during
burner and unit shakedown. This reduced
recommissioning time.
Burner models were 2-D simulations
that used aerodynamics only (no combustion) to represent a single burner in
an idealized tunnel furnace that in turn
represented the equivalent firing wall
The modeling simulated the gas flow inside

the burners in terms of its velocities,
turbulence, temperature, pressure, density,
and local chemical composition.
passages, each of which contains swirl vanes
for spin control. Using shrouds and dampers
to independently control the air ow to each
passage improves control of the stoichiometry at the burner discharge and the burners
ability to reduce NOx emissions as well. The
low-swirl coal spreader disperses the pulverized coal into individual streams that enter
the furnace in a helical flow pattern. The
ames produced are longer than those produced by high-swirl coal spreaders. Finally,
the flame stabilizer ring produces a wellattached, tubular-shaped coal ame, which
further reduces NOx emissions.
Analyze this
The design of the burners and their target
furnace were modeled using computational
fluid dynamics (CFD) techniques. The
modeling simulated the gas ow inside the
burners in terms of its velocities, turbulence, temperature, pressure, density, and
local chemical composition.
Two types of steady-state models were
needed to achieve the project goals. One
modeled the individual burners to ne-tune
region of an outside burner. This

approach yielded a tunnel diameter similar in size to the actual firing environment but without the flame-to-flame
interactions that dominate the flow
behavior several throat diameters from
the wall (Figure 2).
The ow pattern in Figure 2 (L) shows a
well-attached, broad, bushy ame. However, streamlines within the quarl show that
primary air (PA) and secondary air (SA)
streams interact strongly with rapid mixing
driven by the high SA swirl relative to the
axial PA ow. As a result, excess oxygen
enters the primary ignition zone, so NOx
production will be high. The ow patterns
in Figure 2 (R), which are expected to be
enlarged slightly by the high reactivity of
the PRB coal, show a longer, more tubular
ame strongly attached to the burner tip.
The second steady-state model was of
the entire furnace and concentrated on OFA
design and performance. The variables on
which the model focused were upper furnace oxygen distribution, CO and carbon
burnout, the effect of different configura-
2. The streamlines of a CFD-modeled single-burner, pre-retrot (L)

and post-retrot (R)
12.99
12.34
11.69
11.04
10.40
9.75
9.10
8.45
7.80
7.15
6.50
5.85
5.20
4.55
3.80
3.26
2.60
1.96
1.30
0.55
0.00
6.34
6.02
5.71
5.39
5.07
4.75
4.44
4.12
3.80
3.49
3.17
2.85
2.54
2.22
1.90
1.54
1.27
0.95
0.68
0.32
0.00

October 2005
| POWER
tions of mills being out of service on furnace and OFA performance, and possible
approaches for boundary air design to
reduce or eliminate sidewall corrosion,
should it develop.
Because of the left-to-right asymmetry
that results from taking burners out of service, the models of the furnace cover its full
width. Each model uses more than 860,000
brick-like cells to represent the 82-ft-wide
by 51-ft-deep by 184-ft-high furnace volume. More than two-thirds of the cells are
concentrated around the burner openings to
capture reasonable detail of burner flow
patterns, ame structures, and their interactions. Farther away from the burners, the
furnace ow patterns are of a larger scale.
To balance calculation time against ow
detail, the furnace models separated the
pulverized-coal streams to the burners into
three different-sized bins of small, medium,
and large coal particles. This captured most
of the differences in ow behavior among
the variously sized coal particles. The
devolatilization rate and the composition of
volatiles were adjusted for the reactivity of
the particular PRB coal. A two-step reaction system (volatiles to CO to CO 2 )
allowed the realistic simulation of furnace
CO levels from furnace bottom to furnace
exit. Carbon-in-ash (CIA) values were calculated, and small adjustments in the char
burnout reaction rate were made as the base
case simulation progressed. This enabled
the CIA levels of yash at the furnace exit
to remain close to eld data.
Because field-measured values of furnace exit gas temperature (FEGT) and
upper furnace gas temperatures and composition were not available, the base-case
results of the furnace model could not be
validated directly against actual measurements. But it was possible to indirectly validate the model by comparing it to the
model of a pre-retrot furnace with a burner configuration known to produce high
FEGT, as deduced by high steam temperatures and high spray ows.
The FEGT results for the two CFD cases
yielded a difference commensurate with the
back-end heat transfer calculations for the
regular and high-FEGT field setups. No
CFD-based NOx calculations were made.
RPI prefers to rely on regression-analysis
calculations of NO x emissions based on
eld measurements at more than 150 coalred utility units.
In-depth analysis
Figures 3 and 4 show the computed FEGT
and oxygen distribution fields for the preand post-retrot cases. The plots of Figure
3 indicate that the retrot did not create any
49
BURNING PRB COAL

unusual temperature disturbances anywhere
in the furnace and that the incremental
FEGT difference is less than the 50 degrees
F needed to maintain control of ash buildup
in the upper furnace ash. Meanwhile, the
oxygen plots of Figure 4 indicate that,
although the post-retrofit OFA system
made major redistributions of combustion
air and oxygen as part of the NOx-reduction
process, the oxygen distribution in the
upper furnace (and particularly at the furnace exit) is more uniform for the postretrot case.
CFD furnace modeling addressed
potential problems that might surface from
the burner zone air staging that RPI
expected would be required to meet the
projects NO x -reduction goal. It is well
known that opposed-fired utility furnaces
can suffer sidewall corrosion even without
air staging (that is, without OFA), primarily in the middle of the sidewalls at upper
burner elevations. Although sidewall
wastage had not been a problem at this
unit, because burner zone air staging is
part of the NO x -reduction strategy, the
potential for sidewall corrosion may
become more severe in the post-retrofit
situation. Accordingly, a plan to minimize
this potential should be in place. Figure 5
shows the computed gas oxygen content at
the sidewalls. A comparison of Figure 6
(L) to Figure 6 (R) indicates that the postretrofit firing arrangement decreased the
peak sidewall heat flux while increasing
the area of elevated heat flux. This was
due to the spreading out of the combustion process by overfire air.
Last but not least, CFD furnace modeling also helped control project costs.
Another post-retrofit furnace model containing additional wing OFA ports indicated this alternate OFA layout was not as
effective in this unit at upper furnace oxygen mixing and distribution as originally
thought. The CFD modeling informed the
engineering decision to discard this more
complex and costly OFA layout in favor of
the design shown in Figure 4.
Passing with ying colors

The low-NO x burners and OFA were
installed during a scheduled outage of the
plant in the spring of 2002. During the outage, prior to burner installation, the burners
were preassembled on the ground and
hoisted to the burner deck for installation
into the windbox as a one-piece design,
minimizing the construction effort inside
the windbox. They were sized to t within
the existing burner throat openings, avoiding additional modications to the boilers
pressure parts.
50
3. Computed gas temperatures (F) for the pre-retrot (L) and postretrot (R) furnace
3200
3200
2900
2900
2600
2600
2300
2300
2000
2000
1700
1700
1400
1400
1100
1100
800
800
500
500
200
200
4. Computed gas oxygen percentage for the pre-retrot (L) and postretrot (R) furnace
10
10
5. Computed sidewall oxygen percentage for the pre-retrot (L) and

post-retrot (R) furnace
5.0
5.0
4.5
4.5
4.0
4.0
3.5
3.5
3.0
3.0
2.5
2.5
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0

POWER
| October 2005
BURNING PRB COAL

6. Post-retrot sidewall heat ux (Btu/hr-ft2) for pre-retrot (L) and
post-retrot furnaces (R)
Table 2. The results of additional

post-project performance testing
-160000
-160000
-144000
-144000
-128000
-128000
-112000
-112000
Operating
parameter
Gross generation
(MW)
Feedwater flow
(lb/hr)
-96000
-96000
Main steam temp (F)
1,001
1,012
Main steam spray

(lb/hr)
Reheat steam temp
(F)
Reheat steam spray
(lb/hr)
162,000
1,003
1,006
78,000
31,000
-80000
-80000
-64000
-64000
-48000
-48000
-32000
-32000
-16000
-16000
Total air flow (lb/hr)
Average FEGT (F)
Coal flow (lb/hr)
Pre-retrofit Post-retrofit
600
600
4,308,500
4,066,000
664,000
648,000
6,145,000
5,983,000
2,337
2,315
3.1
0.3
3.1
0.158
10
40
0.06
0.09
0.9
Economizer O2 (%)
NOx emissions
(lb/million Btu)
CO emissions
(ppmdv)
Table 1. Comparing test results to project performance targets
Carbon in ash (%)

Operating parameter
Pre-retrofit baseline
Performance target
Performance test
Unit load (MW)
600
600
600
Overfire air (%)
20
14
NOx emissions (lb/million Btu)
0.3
<0.185
0.158
CO emissions (ppm)
10
<100
40
0.06
<2.0
0.09
Unburned carbon in flyash (%)
7. The impact of OFA ow on NOx emissions (in ppm), as a function

of the number of mills out of service (MOOS)
2 MOOS
1 MOOS
NOx measured at economizer out (lb/mmBtu)
0.30
0.25
0.20
0.15
0.10
0
10
15
20
OFA ratio (%)
After the unit was returned to service

and operating close to full load, coal
lines were balanced using variable oriOctober 2005
| POWER
Opacity (%)
fices installed in them during the retrofit. No additional testing for coal fineness was done. Once the coal lines were
balanced, the unit underwent optimization testing and tuning for one week.
CFD facilitated this process as well, by
providing initial starting-point settings
for the burners that were close to the settings determined at completion of the
burner tuning.
As Table 1 indicates, the new burners
met all of the projects performance targets.
Table 2 compares the results of additional
pre- and post-retrofit performance testing
on the boiler at full load. Fossil Energy
Research Corp. (Laguna Hills, Calif.) did
flue-gas emissions measurements and isokinetic flyash sampling at the outlet ducts
of the boilers economizer. Table 2 confirms that the retrofit reduced NOx emissions by nearly 50%.
As expected, NO x emissions decrease
as the OFA ratio increases (Figure 7).
Average NOx emissions of around 0.155
lb/mmBtu, and as low as 0.138 lb/mmBtu
(with an OFA ratio of 20%), were
achieved. The number of coal mills in
operation also had an impact on NO x
emissions. Keeping seven millsrather
than six of the eight availablein service
produced lower NO x emissions because
the low-NOx burners were sized for operation with seven mills. Operating with six
mills produced higher burner velocities
and more turbulent mixing, and higher
NOx as a result.
51

Bps-06 NOx Control

Uploaded by

Copyright:

Available Formats

Bps-06 NOx Control

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bps-06 NOx Control

Uploaded by

Copyright:

Available Formats

BURNING PRB COAL

Minimize NOx using

The burner selected for the project uses

Venturi coal nozzle

Flame stabilizer ring

Secondary air diverter

Secondary air flow

Burner air shroud

Source: Riley Power Inc.

BURNING PRB COAL

their design, to obtain data for the furnace

The modeling simulated the gas flow inside

region of an outside burner. This

2. The streamlines of a CFD-modeled single-burner, pre-retrot (L)

Source: Riley Power Inc.

BURNING PRB COAL

Passing with ying colors

Source: Riley Power Inc.

Source: Riley Power Inc.

5. Computed sidewall oxygen percentage for the pre-retrot (L) and

Source: Riley Power Inc.

BURNING PRB COAL

Table 2. The results of additional

Main steam temp (F)

Main steam spray

Total air flow (lb/hr)

Average FEGT (F)

Coal flow (lb/hr)

Source: Riley Power Inc.

Table 1. Comparing test results to project performance targets

Carbon in ash (%)

Unit load (MW)

Overfire air (%)

NOx emissions (lb/million Btu)

Unburned carbon in flyash (%)

Source: Riley Power Inc.

Source: Riley Power Inc.

7. The impact of OFA ow on NOx emissions (in ppm), as a function

NOx measured at economizer out (lb/mmBtu)

OFA ratio (%)

Source: Riley Power Inc.

After the unit was returned to service

You might also like