Technical Paper

BR-1819

Use of Numerical Modeling to

Compare Overfire Air Systems on
Stoker-Fired Furnaces

Authors:
K.L. Jorgensen
J.R. Strempek
Babcock & Wilcox
Power Generation Group, Inc.
Barberton, Ohio, U.S.A.
W.J. Arvan
Palm Beach Resource Recovery
Corporation
West Palm Beach, Florida
R.H. Schauer
Solid Waste Authority of Palm
Beach County
West Palm Beach, Florida
Presented to:
International Mechanical Engineering Congress and Exposition
Date:
October 31 - November 6, 2008
Location:
Boston, Massachusetts, U.S.A.

Use of Numerical Modeling to Compare Overfire Air

Systems on Stoker-Fired Furnaces
K.L. Jorgensen
J.R. Strempek
Babcock & Wilcox Power Generation Group, Inc.
Barberton, Ohio, U.S.A.

Raymond H. Schauer
Solid Waste Authority of Palm Beach County
West Palm Beach, Florida, U.S.A.

W.J. Arvan
Palm Beach Resource Recovery Corporation
West Palm Beach, Florida, U.S.A.
Presented at:
International Mechanical Engineering
Congress and Exposition
Boston, Massachusetts
October 31 - November 6, 2008

BR-1819

Abstract
The method by which air is introduced to a furnace is
an important aspect of the combustion system. With the
demands of more stringent regulations affecting modern
combustion systems, numerical modeling has become a
valuable tool for design and analysis. Numerical modeling has been used to compare a conventional Controlled
Combustion Zone (CCZ) overfire air (OFA) system design
typically used on stoker-fired furnaces to a more advanced
design. The new design involves removing the opposed
arches and replacing the numerous small overfire air ports
with a smaller number of controlled penetration Precision
Jet ports. Modeling results demonstrate benefits of the
advanced design.

Introduction
The air system is a critical part of the overall combustion system design and an important aspect of system
performance. Often, improvements in boiler performance
and emissions can be realized by air system modifications.
Even more dramatic improvements can be possible with
changes to the basic design of the air system.
The boiler fuel system is the mechanism by which fuel,
the source of thermal energy, is introduced to the boiler and is
the means by which the boiler load is controlled. In contrast,
air systems are used to not only supply oxidant to the fuel,
but as the basic means to control the combustion process
Babcock & Wilcox Power Generation Group

within the overall system. Air systems are routinely used in

an attempt to control the location of combustion by either
retarding or accelerating the local rate of fuel and oxidant
mixing and thus control the rate of heat release. The rate of
mixing also has a large effect on the formation and destruction of carbon monoxide (CO) and nitrogen oxides (NOx).
Much of the successful operation of the overall system relies
on the successful operation of the combustion air system.
Numerical simulation has become an important engineering tool in the design and development of combustion
systems for stoker boilers and furnaces. Using computational
fluid dynamics (CFD) and combustion modeling information, engineers gain important knowledge into the complex
combustion processes that is not available by other means.
Experience and traditional performance predictions are not
as adequate as a model to predict the three-dimensional
distributions of the flow field, gas species, and temperature
throughout a boiler. Field measurements provide actual
performance data but are limited to a few locations from
which interpretations must be made.
Appropriate use of combustion modeling has been recognized as a cost-effective method to evaluate and analyze
multiple design options [1]. This analysis would be cost prohibitive or impossible by a more traditional method, such as
the use of empirical methods and experiments. Additionally,
modeling can be used to speed the design process leading to
shorter, but more thorough, design evaluation.
1

Since 1980, The Babcock & Wilcox Company (B&W)

has invested greatly in the development of its proprietary
COmbustion MOdel. COMO is an advanced, mature design
tool that is routinely used to analyze and assist in the design
of stoker-fired boilers. In addition to waste-to-energy furnaces, COMO modeling software has been proven through
numerous applications such as large-particle ash capture,
Cyclone furnaces, coal gasification, and oxy-combustion,
as well as for power boilers, kraft process recovery boilers,
and the full spectrum of pulverized-coal boilers: tangentiallyfired, single-wall-fired, and opposed-wall-fired. Continued
active development of COMO assures new technologies are
incorporated as they are developed.
The Solid Waste Authority of Palm Beach County owns
two B&W boilers that have been processing municipal
waste for nearly 20 years to reduce the amount reaching the
landfill. While designed to process 730,000 tons of waste
annually, they are currently processing more than 850,000
tons annually. As part of a master plan, these two boilers
are scheduled for refurbishment. Numerical modeling has
been used to incorporate technological advances and design
improvements to enhance combustion control as part of this
refurbishment.
This paper provides a brief description of the capabilities
within COMO as related to waste-to-energy boilers along
with a description of the design and development of a new
air system for this type of furnace.

Combustion modeling
Modeling of solid fuel combustion systems requires
specific treatment of five fundamental processes: fluid
transport, particle transport, homogeneous chemical reactions, heterogeneous chemical reactions, and radiative heat
transfer. The inter-relationship of these five processes is
shown in Figure 1. Only a brief description of the model used
is provided here, however a more comprehensive treatment
can be found elsewhere [2].
Fluid transport Density weighted transport equations
describing the conservation of mass, momentum, component mass, energy, and turbulent parameters are used to
describe fluid flow and heat transfer within the system. The
conservation of momentum and mass are coupled through
the use of the SIMPLE [3] algorithm. Turbulent closure is
accomplished with the application of the widely accepted
k-epsilon model [4].
Particle transport Particle transport can be described
using either an Eulerian or Lagrangian reference frame. The
Eulerian approach [5] assumes that particles move with
the gas flow, while the Lagrangian approach [6] allows
individual particles to follow a path that is coupled with the
flow but not necessarily coincident with it. While Lagrangian
transport can provide a more realistic representation of the
particle flow for larger particles, it is very computationally
expensive. In modeling solid fuel combustion, it is often

Fig. 1 Model for the evaluation of pulverized coal-fired combustion based upon five fundamental processes.
2

Babcock & Wilcox Power Generation Group

beneficial to employ a combination of both of these approaches due to the wide distribution of particle sizes. This
is accomplished by using the efficient Eulerian reference
frame for small particles, which travel with the gas, and a
Lagrangian reference frame for larger particles, which may
not flow with the gas.
The effect of fluid turbulence on particle transport is
modeled using the Lagrangian stochastic deterministic
model described by Milojevic [7]. This model utilizes the
fluctuating velocity component in addition to the mean
velocity to determine the influence of the gas flow field on
particle motion.
Homogeneous chemical reactions Gas phase
combustion involves the reaction of various gas species.
A conservation equation is solved for each gas species being modeled. An appropriate combustion model is used to
determine the mean production rate for each of the species
involved.
While various methods can be used to determine the
production rate of each species, two approaches are commonly used in the present model. The first is known as the
eddy dissipation combustion model (EDM) developed by
Magnussen and Hjertager[8]. The second approach, known
as the eddy dissipation concept (EDC), was also developed
by Magnussen [9]. The rate of combustion involves a single
step in the EDM approach which is controlled by the turbulent mixing of reactants.
In practice, the combustion process can be further limited
by including a second step to the process which involves the
moist oxidation of CO:
CO + (1/2) O2 CO2
The combustion rate for this step is limited by the minimum of the global reaction kinetic rate [10] and the rate of
turbulent mixing [8].
The second approach proposed by Magnussen [9] is the
eddy dissipation concept (EDC) which overcomes the limitation of EDM where the reaction rate is controlled simply by
turbulent mixing. The EDC model can accommodate kinetic
limited reactions such as premixed combustion and the individual rates for each reaction in a particular mechanism. The
EDC model is considerably more computationally intensive
than the EDM model and is not commonly used on routine
boiler modeling efforts.
NOx model Two methods can be used to predict the
formation and reduction of nitrogen oxides (NOx). The first
is a global NOx model used with the EDM combustion
model. The global NOx model contained in COMO can
predict both the formation and reduction of NOx. During
devolatilization, fuel-bound nitrogen is assumed to form
HCN and NH3. These species react with the flue gas and are
either oxidized to form NOx or reduced to form N2 [11]. The
formation and destruction of thermal NOx is described by
the modified Zeldovich mechanism [12]. Nitrogen released
from the burning of char is assumed to form N2 and NO.

Babcock & Wilcox Power Generation Group

Because the concentrations of NOx pollutants are typically

several orders of magnitude lower than that of the other
combustion species, the emission prediction mechanism can
be de-coupled from the combustion solution and solved after
the combustion solution has been converged.
The second method is a detailed mechanism that includes
the various nitrogen species and their reactions in the context
of the EDC model described previously.
Heterogeneous chemical reactions Solid fuel
combustion is a complex process involving drying, devolatilization and char oxidation. These processes are modeled
by three steps or global chemical reactions:
Wet Fuel Dry Fuel + H2O
Dry Fuel Volatile Gas + Char
Char + O2 CO + CO2 + H2O + Ash
As the fuel particle absorbs heat by radiation and convection, the moisture in the particle evaporates, leaving
dried fuel. A further increase in the fuel temperature causes
pyrolytic reactions that form volatile gases and char. The
volatile composition depends on several factors including
fuel elemental composition and environment, but typically
includes species such as CH4, SO2, CO2, and H2O.
Devolatilization in the present model can be represented by two first-order kinetic expressions described by
Ubhayaker et al. [13].
Following devolatilization, the remainder of the particle
consists of char residue and inert ash. The char is assumed to
react heterogeneously with the oxidizer. The basic approach
is described by Field et al.[14]. The carbon burnout kinetic
(CBK) [15] model developed by Hurt et al. has also been
implemented in the current model.
Radiative heat transfer The well known discrete ordinates method (DOM) described by Fiveland [16] is used to
solve the radiative transport equation for a specified number
of ordinate directions. An appropriate model, such as the
wide band (Edwards, 1976) or the weighted-sum-of-graygases, is used to determine radiative transport properties of
the fluid. The effect of particles on the radiative transport
is also accounted for through absorption, emission and
scattering. Mie Theory is used to determine the scattering
coefficient.
Grate combustion Fuel particles that land on the
grate are treated together in a separate surface combustion
model. In the model, all moisture and volatiles are driven
off and released to the fluid domain above the grate. Char
oxidation is based on grate stoichiometry. The amount of CO
and CO2 generated from the reaction is strictly a function
of the oxygen available from the undergrate air stream. If
there is insufficient oxygen to convert all of the carbon to
CO, then some carbon will be lost as unburned on the grate.
The reaction products are released to the fluid domain in a
similar manner as the moisture and volatiles.

Stoker boiler air system development

A new air system has been developed for use on original
equipment and retrofit stoker boilers. This new design was
developed to improve performance of the common stoker
boiler air systems. These older arrangements incorporated up
to eight separate levels of overfire air (OFA) with each level
consisting of a large number of small ports. As an example,
the boiler described by Dudek [17] consisted of four levels
of OFA with a total of 15 inlet nozzles on each level.
The new design consists of two levels of OFA and straight
furnace walls. The portion of OFA is similar to original
designs. To accommodate the increased air flow through a
reduced number of ports, larger ports were utilized.
Waste-to-energy retrofit The example presented here
shows the retrofit of a waste-to-energy (WTE) boiler that was
originally designed with the B&W Controlled Combustion
Zone (CCZ) furnace. However, this new design is applicable
on any stoker boiler whether as a retrofit or as original
equipment. Major differences between the original design
and the new design are compared in Table 1. An outline of
the original and new design in shown in Figure 2.
Gas flow in lower furnace The large number of small
air nozzles and the multi-level characteristic of the original
design were meant to provide flexibility in the distribution
of OFA. In addition, the two opposing wall arches define
a combustion zone. The large number of small air nozzles
limits the amount of air that can be introduced through a
single nozzle which, in turn, limits the air jets penetration
into the furnace. This could lead to less than adequate mixing
with the gas stream. The two opposing arches help alleviate
this by creating a smaller cross-section for mixing to occur,
however the flow is constrained to the center of the furnace.
The result of less than adequate penetration and the presence
of the arches is that the flow entering the upper furnace is
concentrated in the center where the vertical velocity and
gas temperature are high. This creates large velocity and
temperature gradients in the gas stream entering the convection section of the furnace.
The resulting effect of the arches on the flow in the furnace can be clearly seen in Figure 3(a). This figure shows
the gas velocity in the lower furnace and just above the CCZ
arches. Here, it is clear that the jets from inlets A, B and the
wind swept spouts (WSS) do not penetrate to the center of
the furnace. It is also clear that the jets from inlets C through
F are able to just penetrate to the center of the furnace at its
narrowest point. Because the flow is restricted through this
Table 1
Comparison of Original and New Air
System Configurations
Wall Design
Levels of OFA
Total No. of Ports

Original Design

New Design

CCZ

Straight

182

14
(3-front/2-rear
5-front/4-rear)

Fig. 2 Boiler outline of original and new design.

region, all of the combustion gases are concentrated into a

single column of gases in the center of the furnace.
The space between this central column of combustion
gases and the front and rear wall has relatively little gas
movement. In some locations, large recirculation zones
develop. The net result of this flow pattern is that only a
limited portion of the furnace volume immediately above
the CCZ arches is being fully utilized.
As this column of combustion gas leaves the lower furnace and enters the tube banks, the flow is somewhat dissipated through the tube banks. However, at the point where
the gas enters the tube bank, a region exists where the flow is
still at a much higher velocity and a higher temperature than
the surrounding gas. The combined effect of high velocity
and temperature can lead to accelerated tube wastage.
Figure 3(b) shows the flow in the lower furnace of the
new design. The effect of removing the CCZ arches is immediately evident. Two significant characteristics can be
seen: 1) the new OFA jets penetrate completely across the
furnace; 2) the flow below the lower OFA ports (zone 1)
is much slower and more uniform than the flow below the
CCZ arches.
Because the OFA jets penetrate the complete depth of
the furnace, flow within the furnace can be more easily
controlled. The improved control leads to increased utilization of the furnace volume. It is evident that the new air
system produces a much more uniform gas flow throughout
the furnace. The more uniform gas flow above the upper
OFA level is expected to reduce the rate of corrosion in the
superheater.
It is also expected that the lower gas velocities above the
stoker will help control grate combustion and carryover.
Combustion characteristics The original design
introduced all of the fuel and air in the region under the two
opposed wall arches where the majority of the combustion
took place. The rest of the furnace completed the combustion process in a burnout region. Presence of the arches
prevented much flexibility in controlling where combustion
took place.
Babcock & Wilcox Power Generation Group

Fig. 3 Velocity vectors in the lower furnace for (a) original design and (b) new design.

In contrast, the new design spreads out the combustion

through two zones (Figure 3) and significantly reduces the
burnout region. Because the amount of air that is delivered
to the furnace through the upper and lower OFA levels can
be controlled, this new design provides much greater control
of the combustion process.
Figure 4 shows the average gas temperature as a function
of furnace height. The peak gas temperature for the original
design occurs below the arches where the majority of the
heat is being liberated. Above the arches, the gas temperature
slowly decreases through the remainder of the furnace.
In contrast, the peak temperature for the new design occurs above the lower OFA ports. While there is an increase
in temperature between the stoker and the lower OFA ports,
there is not enough oxygen to complete the combustion
process. The lower OFA ports provide nearly all the oxygen
required to complete the combustion process which accounts
for the elevated temperatures between the two OFA levels.
The remainder of the oxygen is supplied by the upper OFA
ports which allows the combustion process to be completed
and also has a tempering effect on the combustion gases.
Above the upper OFA ports, the combustion gases behave
in a similar fashion as the original design. The delayed heat
release in the new arrangement accounts for the slightly
higher temperatures in the upper furnace.
The original design was very effective at completing
the combustion process below the arches through a rapid
mixing of the fuel and air. One concern with removing
the arches was that mixing of the air and fuel may not be
sufficient to complete the combustion process before the
Babcock & Wilcox Power Generation Group

furnace exit. Carbon monoxide predictions can be used as

a method to compare the mixing effectiveness between the
two designs.
Figure 5 shows the CO concentration as a function of
furnace height. Concentration of CO in the original design
decreases rapidly below the arches because all of the air
required for combustion is introduced into the furnace in
this region. Above the arches, CO concentration gradually
decreases as the gases pass through the burnout zone.
The CO concentration for the new design only gradually
decreases from the stoker to the upper OFA level. Below
this point in the furnace, there is not enough oxygen to
complete the combustion process and CO concentrations
cannot decrease any further. Above the upper OFA level,
the CO concentration decreases rapidly. This decrease is
similar to the rapid decrease below the arches in the original
design. From this figure, it is clear that the new OFA ports
provide the mixing required to complete the combustion
process. The two different methods used for mixing are also
illustrated in Figure 3.
NOx control Another benefit of the new design is the
improved NOx emissions characteristics of the furnace.
Model predictions indicate that the new design provides a
reasonable reduction in the base NOx. Figure 6 provides a
comparison of the original and new designs. The multistage
combustion processes provide two distinct zones where the
combustion gases are in a substoichiometric condition. This
slows the production of NOx while promoting the reduction
of NOx. The original design lacks any such reduction zone.
By reducing the base NOx, this new design will reduce the
5

Fig. 4 Temperature as a function of furnace height.

Fig. 5 Carbon monoxide as a function of furnace height.

cost associated with controlling NOx to future compliance

levels.

furnace model. The benefits of the new design, when compared with the existing arrangement, have been presented.
The knowledge gained in developing this new design will
continue to be useful in designing and developing other new
boilers and retrofits on existing units.

Conclusions
Since 1980, The Babcock & Wilcox Company has spent
considerable resources developing and maintaining a stateof-the-art comprehensive combustion model. Recent use of
this model has demonstrated how modeling can be used in
furnace design and development.
The next generation in stoker furnace air system design
has been developed with the use of this comprehensive

Acknowledgements
The authors would like to thank Scott A. Dudek and
William J. Kahle for their contribution and assistance during the preparation and interpretation of a portion of these
results.

Babcock & Wilcox Power Generation Group

Fig. 6 Comparison of NOx concentration for (a) original design and (b) new design.

Babcock & Wilcox Power Generation Group

References
[1]



[2]




[3]


[4]



[5]




[6]



[7]




[8]




[9]

Fiveland, W.A., 1996, Combustion Modeling

for Fossil-Fired Applications, 1996 National
Heat Transfer Conference, Houston, TX,
August 3-6.
Steam/its generation and use, 41st Edition,
Chapter 6: Numerical Modeling for Fluid Flow,
Heat Transfer, and Combustion, 2005,
The Babcock & Wilcox Company, Ed. S.C. Stultz
and J.B. Kitto.
Patankar, S.V., 1980, Numerical Heat Transfer
and Fluid Flow, Hemisphere Publishing Corpora
tion, New York.
Launder, B.E., and Spalding, D.B., 1974, The
Numerical Computation of Turbulent Flows,
Computer Methods in Applied Mechanics and
Engineering, 3.
C r o w e , C . T. , a n d S m o o t , L . D . ,
Multicomponent Conservation Equations,
Pulverized Coal Combustion and Gasification,
edited by L.D. Smoot and D.T. Pratt, Plenum
Press, New York, 1979.
Crowe, C.T., Gas-Particle Flow, Pulverized
Coal Combustion and Gasification, edited by
L.D. Smoot and D.T. Pratt, Plenum Press,
New York, 1979.
Milojevic, D., 1990, Lagrangian Stochastic-
Deterministic (LSD) Predictions of Particle
D i s p e r s i o n i n Tu r b u l e n c e , J o u r n a l o f
Particles and Particle Systems Characterization, 7,
pp. 181-190.
Magnussen, B. F. and Hjertager, B. H., 1976,
On Mathematical Modeling of Turbulent
Combustion, 16th Symposium (International)
on Combustion, The Combustion Institute,
Pittsburgh.
Magnussen, B. F., 1981, On the Structure
of Turbulence and the Generalized Eddy
Dissipation concept for Turbulent Reactive
Flows, Proceedings of the 19th American
Institute o f Aeronautics and Astronautics
Aerospace Science Meeting, St. Louis, MO.

[10] Howard, J.B., Williams, G.C., and Fine, D.H.,

Kinetics of Carbon Monoxide Oxidation in
P o s t f l a m e G a s e s , 1 4 t h S y m p o s i u m
(International) on Combustion,The Combustion Institute,
Pittsburgh, pp. 975-986, 1972.
[11] de Sote, G.G., Overall Reaction Rates of
NO and N 2 Formation from Fuel Nitrogen,
15th International Symposium on Combustion,
T h e C o m b u s t i o n I n s t i t u t e , P i t t s b u rg h ,
p. 1093, 1975.
[12] Bowman, C.T., Kinetics of Pollutant Formation
and Destruction in Combustion, Progress in
Energy and Combustion Science, Vol. 1,
pp. 33-45, 1975.
[13] Ubhayakar, S.K., Stickler, D.B., Von Rosenberg,
C.W., et al., 1975, Rapid Devolatilization of
Pulverized Coal in Hot Combustion Gases,
16th Symposium (International) on Combustion,
The CombustionInstitute, Pittsburgh.
[14] Field, M.A., Grill, D.W., Morgan, B.B., et al.,
1967, Combustion of Pulverized Coal, The
British Coal Utilization Research Association,
Leatherhead, Surrey, England, United Kingdom.
[15] Hurt, R., Sun, J.K., Lunden, L., 1998, A Kinetic
Model of Carbon Burnout in Pulverized Coal
Combustion, Combustion and Flame, Vol. 113,
pp. 181-197.
[16] Fiveland, W. A., 1984, Discrete-ordinates
solutions of the radiative transport equations
for rectangular enclosures, Transactions of
American Society of Mechanical Engineers
Journal of Heat Transfer, 106, pp. 699-706.
[17] Dudek, S.A., Wessel, R.A., Strempek, J.R., 1999,
Three-Dimensional Numerical Modeling
of Stoker-Fired Power Boilers, ASME
International Mechanical Engineering Congress
and Exposition, Nashville, TN, November 14-19.

Copyright 2008 by Babcock & Wilcox Power Generation Group, Inc.

a Babcock & Wilcox company
All rights reserved.
No part of this work may be published, translated or reproduced in any form or by any means, or incorporated
into any information retrieval system, without the written permission of the copyright holder. Permission requests should be addressed to: Marketing Communications, Babcock & Wilcox Power Generation Group, P.O.
Box 351, Barberton, Ohio, U.S.A. 44203-0351. Or, contact us from our Web site at www.babcock.com.

Disclaimer

Although the information presented in this work is believed to be reliable, this work is published with the
understanding that Babcock & Wilcox Power Generation Group and the authors are supplying general information and are not attempting to render or provide engineering or professional services. Neither Babcock &
Wilcox Power Generation Group nor any of its employees make any warranty, guarantee, or representation,
whether expressed or implied, with respect to the accuracy, completeness or usefulness of any information,
product, process or apparatus discussed in this work; and neither Babcock & Wilcox Power Generation Group
nor any of its employees shall be liable for any losses or damages with respect to or resulting from the use of,
or the inability to use, any information, product, process or apparatus discussed in this work.