Improved Low NOx Firing Systems for Pulverized Coal Combustion

K. McCarthy, S. Laux, J. Grusha

Foster Wheeler Energy Corporation
Clinton, NJ 08809-4000, USA
T. Rosin
TR-Tech International Oy
Turku, Finland
G. L. Hausman
PP&L, Inc., Martins Creek SES
Martins Creek, PA, USA

Abstract
More stringent emission limits or the addition of post combustion NOx control create the
need for improvements of NOx emissions from pulverized coal boilers. Many boilers
retrofitted with Low NOx technology during Phase I and Phase II of the CAAA fail or
marginally meet their requirements.
Technical solutions range from addition of overfire air and state-of-the-art low NOx
burners to low cost additions of combustion enhancements. Regardless of the
combustion NOx control method used, stoichiometries local to the burners must be
maintained at the designed values at all times to provide high NOx performance at low
efficiency loss due to unburned fuel.
This paper describes Foster Wheelers approach to NOx emission improvements for
existing low NOx firing systems. The technology to measure air and coal flow
individually for each burner and to control the parameters for optimum combustion are
presented and discussed. Field experience shows the installation and advantages of
the technology.

The Need for Better Air and Fuel Balance

Low NOx firing systems were installed in Phase I and II of the Clean Air Act
Amendment to meet federally mandated NOx emission limits. Although Phase II is
nearing completion by the year 2000, tighter emission control is mandated for 22
eastern states as part of EPAs recent ozone regulations. With a few exceptions, the
NOx limits for ozone control during the summer months require a post-combustion
control method such as SCR on coal fired boilers, because Low NOx firing systems
alone can not achieve compliance. However, investment costs and ammonia costs of
an SCR DeNOx catalyst can be greatly reduced, if the output NOx from the firing
system is low. Thus, more stringent emission limits and the addition of post combustion
NOx control create the need for improvements of NOx emissions from the firing system
of pulverized coal boilers.
Even boilers retrofitted with Low NOx technology during Phase I and Phase II of the
CAAA are required to further reduce emissions for compliance margin or total cost
considerations. Currently, Low NOx firing systems are installed in almost all US boilers.
Users experience performance degradation with time from the initial retrofit which
lowers the margin to achieve the federal limits. Severe cases threaten to de-rate the
boiler capacity. Some utilities require to gain additional NOx reduction from the existing
system, because focus has changed from pure NOx compliance to NOx bubbling.
Depending on the target NOx, improvement of existing Low NOx firing systems can
comprise of an OFA retrofit, advanced LNB or a combination of these. For many boilers
this might not be cost effective and performance improvement of the systems already
installed would be the best choice.
Regardless what the NOx reduction approach, all low NOx firing systems depend on a
defined balance of the air and coal at the burners. By monitoring and controlling the air
and coal flow to the burners, existing as well as new systems greatly benefit from lower
emissions and improved boiler operation. Or, if you use the following analogy for boiler
emission improvement: This trend is comparable to the change from carburetors to fuel
injection systems seen over time in the car industry.
Almost every pulverized coal boiler experiences some of the problems associated with
poor air and fuel distribution:

Poor emission performance

Increased unburned carbon in the fly ash
Distorted oxygen profile at boiler outlet
Flame impingement
Increased slagging
Water wall wastage

Especially small single wall fired units suffer from these problems. But these problems
are not limited to one firing system configuration or size. In the past, industry believed
that tangentially fired furnaces were not sensitive to the air and coal balance. Recent
investigations show the benefits of balancing of T-fired units as well.

CO / LOI
Flame Impingement

NOx

High Air
Low Fuel

Slagging/Water Wall Corrosion

Any deviation from the design flow values at individual burners result in some burners
operating at a fuel lean or a fuel rich condition. The fuel rich burner is producing large
amounts of CO, high LOI and longer flames while locally lowering the oxygen level in
the flue gas (Figure 1). On the other hand, the fuel lean burner produces high NOx
levels at elevated O2. The outcome at the boiler exit is a flue gas with high CO and high
NOx. In addition, LOI is elevated due to the burners operating at low stoichiometries.

Low Air
High Fuel

Figure 1: Results of air and fuel imbalances

Ideal match of air and fuel at all burners is a theoretical assumption that is not validated
by every day realities in coal fired power plants. Windbox and air duct designs as well
as location and different swirl settings of the burners prevent air flow to be the same to
all burners. The design of the pulverizers and the different length and layout of the coal
conduits result in an uneven fuel distribution at the burners which also changes with
unit load.
These imbalances have been addressed in the past by optimizing the windbox air
distribution and adding orifices to improve the fuel distribution. Optimizations are
usually very time consuming and require physical and/or numerical models. Some
vendors recommend burner modifications at the coal nozzle or in the secondary air to
influence maldistribution. However, these attempts fall short of fixing the fundamental
problems and are merely cosmetic.

Most Low NOx Burners allow to control the air flow through movable sleeve dampers or
other designs. These are adjusted mainly during initial optimization of the Low NOx
system after a retrofit. Usually, it was not necessary to address distribution changes
with load and degradation of system performance with time. In the recent past, several
vendors developed control strategies and software to improve boiler and emission
performance. While this is the step needed for improvement, most of the systems lack
reliable flow measurement and control of air and fuel.
Foster Wheeler Energy Corporations field tests have confirmed that improvement in air
and fuel balance reduces NOx and CO emission levels at the same time. Higher than to
be expected LOI for a given boiler size and fuel fineness can usually be attributed to
individual burners badly out of balance. Balancing across all burners has the following
benefits:

Meet NOx emission with sufficient margin

Lower LOI
Lower slagging
Margin to lower excess air
lower NOx
less auxiliary power
higher efficiency

Air Balance
In most US wall fired boilers, a windbox common to all burners on one wall is used to
supply the air necessary for the combustion process. On some boilers compartmented
windboxes allow air bias to certain elevations of the firing system. In contrast to
individual burner air supply seen often outside the US, the windbox design results in
lower investment costs but has the inherent problem that the air is not distributed
evenly to all burners. This is due to different local differential pressure between the
windbox and the furnace as well as duct design.
To mitigate the air mal-distribution, burners are designed with perforated plates at the
register inlet and with movable flow control devices such as sleeve dampers. These are
normally moved to defined positions for burner out-of-service, light-off and operation.
The operating position is set individually in accordance to the results of a balanced
oxygen profile at the back end of the boiler during optimization. However, most drives
for the air control devices are capable to assume any position between the end stops
and are designed for modulating operation. Thus, it is possible on most burners to
control the air flow once it has been measured.
Foster Wheeler uses commercially available burner air measurement to determine the
secondary air flow (See Figure 2). These devices are mounted into the swirling air of
the secondary air annulus of the burners. The probe measures the velocity head of the

flow largely independent of the direction of the swirling flow. The flow rate is then
determined by correcting the measured velocity head with the temperature of the
windbox air and the direction of the swirl vanes. Appropriate dampening and averaging
of the signal is necessary to use the output for a closed loop control of the desired flow
rate by the sleeve damper position.
OUTER SECONDARY
AIR PROBE

STAINLESS STEEL
TUBES TO MAG GAGE
OR TRANSMITTER

Figure 2: Burner air measurement installed in the outer secondary air annulus of a
Foster Wheeler burner
Figure 3 shows the principal dependency of the pressures measured by the IBAM as a
function of the angles of the swirl vanes and the burner flow rate for a Foster Wheeler
Controlled Flow dual register. It can be seen that the signal for the flow rate varies with
the different positions of the outer register.

Measured Velocity Head

Dependency of Throat Velocity Head on

Register Vane Angle and Flow Rate

Burner Air Flow Rate

Increased Swirl

20

40

60

80

100

Vane Angle (deg)

Figure 3: General change of air flow probe pressures with register vane angle and
burner flow rate

Fuel Balance
On coal fired boilers it is difficult to determine the fuel flow to individual burners, since
coal particles are transported in a gas suspension that is governed by the complex
physics of two-phase flow. Total fuel flow to the pulverizer is usually known from the
gravimetric feeders and air flow into the pulverizer is measured by flow venturies or
other methods. The flow of air and coal can be easily controlled by damper position and
speed control of the feeders, respectively. However, the distribution of the usually
inhomogeneous particle suspension is difficult and depends on a lot of variables and is
different from pulverizer to pulverizer.
Measurement of suspended particle flows is even more difficult. Various extractive
methods are performed manually (ASME method, RotoProbe, etc.) and can give only a
snapshot result of fuel distribution. Fuel roping and changes with load are major
concerns with regard to the accuracy of these methods. Thus, these methods were
used mainly for initial boiler optimization and pulverizer performance checks.
A continuous measurement of the fuel flow in each conduit is required for on-line
optimization of the air and coal balance at each burner. Currently, the Electric Power
Research Institute is conducting a study of commercially available systems to
determine the coal flow in the conduit leading to the burners. Among them are
microwave, acoustic, and electro-static devices.

Electric Charge Transfer Technology

Foster Wheeler is partnering with TR-Tech International Oy of Finland to market the
Electric Charge Transfer technology (ECT) worldwide. The patented technology
measures the electric charges present in any two phase flow transport and uses the
signals to determine the relative coal distribution between the conduits of one mill. The
absolute flow in each conduit can be readily calculated from the total coal flow to the
mill.

ECT Burner Coal Flow Measurement

Relative coal flow measurement between
conduits
Continuous, online
Measures electric charges present
in any two-phase flow application
due to relative motion

+
+

HARDENED
ANTENNAS

COAL
CONDUIT

COAL
PARTICLES

Figure 4: Major components of the Electric Charge Transfer system (ECT)

The ECT system consists of three receiving antennas in each coal conduit that are
connected to a signal conditioning unit. This signal conditioning unit is in turn
connected to a personal computer that is used for data processing and analysis. TRTechs proprietary software is used to determine the balance between the conduits of
one mill, to display the results to the operator and to feed the data to an optional
connection to the plants DCS system for continuos optimization.

The antennas are easily installed through the wall of the existing conduit and they are
inserted approximately one inch into the coal stream. Three antennas in one conduit
are connected together. Their location in the pipe wall was determined so that the
effects of coal ropes on the measurement results are minimized. Antennas are made of
hardened steel to ensure long operating life. The installation is very simple and
requires only a mill taken out of service for several hours which minimizes lost
generating capacity. After installation, the ECT measurement is verified by a standard
ASME or other conventional sampling procedures and the system is then ready to be
used.
The ECT system offers several distinct advantages:

On-line and continuous information about the relative distribution of the fuel
between the conduits.
ECT is not effected by coal type, moisture, ash content or coal roping.
The signal condition unit can be as far as 1200 feet from the conduits.
The abrasion resistant antennas in the coal conduit are passive, no power supply
is necessary for them.
The method is very cost effective, since mostly standard equipment is used.
Inexpensive coaxial cable connects the antennas to the signal conditioning unit.
The data collection and processing is done on a standard personal computer.
The installation is easy and can be done during short mill outages.

Figure 5: ECT system equipment with antennas

Field Verification of the ECT Technology

ECT systems have been successfully installed on 3 limestone plants and 5 utility steam
generators. Recent US field verifications test were performed by Foster Wheeler and
PP&L, Inc., at Martins Creek Steam Electric Station. Unit 1 is a 120 MW Foster
Wheeler front fired boiler with two ball tube mills serving a total of 12 burners. At each
end of a mill an exhauster and a three way riffle distributor are used to feed the fuel to
the conduits. The ECT system was installed on all six coal conduits of the second mill
immediately below the coal sampling ports.
The following graphs compare the results of the standard ASME test with those
obtained with the ECT method. The ECT values were from two time averaged periods
at the beginning and the end of the ASME test. For the first test, the exhausters at both
end of the mill were loaded evenly which should yield an ideal mass distribution of 16.7
% for all conduits. Figure 6 shows results of this test.
Comparison of ECT Results to ASME Test
No End to End Mill Bias
Mass Distribution [%]

30
25

ASME
E C T average

20
15
10
5
0
1

3
4
C o nduit

Figure 6: Comparison of ASME coal balance test with ECT Method for all six conduits
loaded evenly
The results of both measurement methods compare very well. The maximum deviation
between the two test methods is less than 1% of the mass distribution. In addition, the
test shows that conduit 3 and 4 are loaded approximately 2 and 3% higher than the
average, respectively. Expressed as a percentage of the average coal flow these
conduits have 12 and 17% higher coal flow. Considering the marginal performance of
three way distributors this result is satisfactory.
In a second test, the mill output was biased by the exhausters towards conduits 4
through 6. Thus, the ideal distribution in this case would be 10 % mass distribution for
conduits 1 to 3 and 23.3% for the conduits 4 to 6. Figure 7 shows the results of this
test.
9

Comparison of ECT Results to ASME Test

30%/70% End to End Mill Bias
Mass Distribution [%]

30
25

ASME
E C T average

20
15
10
5
0
1

3
4
C o nduit

Figure 7: Comparison of ASME coal balance test with ECT Method for flow biased to
three of six conduits
With the exception to conduit 4 the difference between the ASME method and the ECT
method is again approximately 1 % mass distribution or less. Only conduit 4 shows a
large difference of 3 % between the two methods. This conduit being normally highly
loaded is probably close to a choking condition when it is operated with approximately
50% higher coal flow. The unsteady effects can not be measured with the short term
ASME method, but are recorded with the ECT antennas. This experiment with extreme
mill bias towards one end of the mill also shows that the flow is not biased as expected.
Instead of the theoretical value of an end-to end bias of 30/70%, only approximately
40/60% were achieved.
In general, the results of the ECT method compare very well with a standard ASME
probe test even at extreme shifts of the coal flow between the conduits. This confirms
the viability of the technology for real time coal balancing applications.

Control of Coal Distribution

The ECT measurement technology provides the backbone for the balancing of the fuel
flow. Its on-line information can be forwarded to the plants Distributed Control System
for necessary balancing of the fuel flow.
An initial and very cost effective method can be to just measure the fuel imbalances in
the system and use the information to calculate the correct air flow to each burner. As
discussed previously, the secondary air flow to the burners can be controlled by the
sleeve dampers of the burners. Although the fuel imbalance is not improved, each
burner operates at the design stoichiometry with optimum NOx and combustion
performance. Depending on actual fuel flow to the burner, each burner has a slightly
10

different load which is not considered to be a problem in a multi-burner boiler. Since the
adjustment is made on the air side, this performance enhancement scheme is limited to
cases where the fuel imbalance is less severe and windbox pressure is not design
limited.
Several devices are available to directly influence the distribution of the coal. The best
choice depends on the pulverizer design, space available and on the coal conduit
layout. Thus, only a few examples can be given here.
Orificing the coal conduits has been done in the past with fixed orifices to improve the
fuel distribution. By using adjustable orifices with actuators the fuel distribution can
influenced by the DCS and set to the target values. This method has a serious
disadvantage, however. An orifice creates additional pressure drop and mainly
influences the flow of the transport air in the conduit. Due to changes in pressure drop
the particle distribution is only changed indirectly. This can result in operation with
insufficient transport air flow and consequently the danger of coal layout in horizontal
pipe runs and in the burner. Orificing is not a risk free, standard solution to coal
distribution problems.
Improved fuel distribution can be achieved by adjustable riffle boxes. These riffle boxes
allow to bias the particle flow towards individual conduits. The design is similar to the
fixed rifflers, but turning vanes are used at the riffler inlet. The change of a fuel system
without distributors to this design requires additional space above the pulverizers which
might be limited in retrofit applications.
Dynamic classifiers offer not only improved particle fineness but have also by design a
much better conduit to conduit balance. As part of a major boiler improvement a retrofit
with dynamic classifiers might be economical. Design features include distribution
vanes at the classifier outlet to direct the particle flow towards specific conduits. These
vanes can be equipped with an actuator for on-line fuel flow control to the burners.

Conclusions
Well balanced air and fuel flow at the burner is needed to get the highest performance
from existing or new Low NOx firing systems. Developments to measure the air and fuel
distribution continuously at the burners are the backbone for automated systems which
constantly monitor and adjust the stoichiometry at the burners. Because of the complex
nature of coal flow in a conduit it was not possible to measure the coal flow on-line until
recently. TR-Tech and Foster Wheeler Energy Corporation are continuing to develop
and market the ECT system for continuos coal flow measurement. ECT was developed
to provide coal fired boiler operators with the capability of getting additional NOx, CO
and LOI reduction from their combustion systems.

11