Fuel 82 (2003) 1127–1137

www.fuelfirst.com

Further study of the gas temperature deviation in large-scale

tangentially coal-fired boilersq
Chungen Yin*, Lasse Rosendahl, Thomas J. Condra
Institute of Energy Technology, Aalborg University, Pon 101, DK-9220 Aalborg East, Denmark
Received 21 February 2002; revised 18 December 2002; accepted 18 December 2002; available online 17 January 2003

Abstract
Gas temperature deviation in upper furnace is an important but a less reported issue in large-scale tangentially fired boilers, since they
endanger largely boilers operation. Simulations are conducted in this paper to study the deviation. Perfect agreement between the simulation
results and key boiler design values and available site operation records indicates that the calculations are reliable. Based on the simulations,
effect of some factors, including residual airflow swirling at furnace exit, super-heaters panels, coal particle trajectories and their combustion
histories, on temperature deviations are studied in details. The most important cause and how it affects the temperature deviation are located.
Two new methods, a nose on front-wall and re-arranged super-heater panels, are put forward unprecedentedly to alleviate the deviations.
q 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Tangentially fired; Pulverized coal; Boiler; Gas deviation; Simulation

1. Introduction believed that they are related to steam flow deviation among
panels on the steam side and uneven gas temperature
Tangentially fired boiler is the most commonly used distribution in upper furnace and in crossover pass.
industrial coal combustion mode and, thus gains much As a special topic, the gas temperature deviation in large-
attention. With the development of comprehensive com- scale tangentially fired boilers was studied [12]. The
bustion models and their successful applications in modeled gas temperature deviations were compared with
industrial coal-fired boilers, as surveyed in Refs. [1,2], the measured values in a 600 MW boiler, and an
the complex phenomena in tangentially fired boilers, approximate agreement between calculations and measure-
including gas – solid flow, combustion, heat transfer and ments was found. It was concluded that residual airflow
NOx reduction, are widely studied with simulations and swirling and platen super-heaters were the two important
better understandings are achieved [3 –8]. In some cases, it causes for the gas temperature deviation.
replaces physical experiments with equivalent ‘numerical’ Following in the wake of earlier studies, this paper is
experiments. More recently, Fan et al. modeled system- devoted to investigating in details the effects of some
atically the combustion processes in a large-scale tangen- factors, including residual airflow swirling at furnace exit,
tial boiler (600 MW), including the flow, combustion and existence of super-heaters panels, coal particle trajectories
NOx emissions, and used their achievements to improve and their combustion histories inside and beyond combus-
boilers’ performance [9 –11]. tion section, on temperature deviations in tangentially fired
Compared with the above achievements, studies on gas boilers, based on which possible measures are put forward
temperature deviations in upper furnace in tangentially fired to alleviate the deviations.
boilers are less reported. In fact, it is also an important issue.
Tube overheating and explosion of super-heaters and/or re-
heaters rank in the first place among all the operational 2. The 609 MW utility boiler
accidents of large-scale tangentially fired boilers. It is

* Corresponding author. Tel.: þ 45-9635-9248; fax: þ 45-9815-1411. The 609 MW boiler is shown schematically in Fig. 1.
E-mail address: [email protected] (C. Yin). The lines, P1, P2, P3, P4, V1, V2 and V3, denote some
q
Published first on the web via Fuelfirst.com—http://www.fuelfirst.com cross-sections in upper part of the boiler. The dots L0 – L4
0016-2361/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0016-2361(02)00418-0
1128 C. Yin et al. / Fuel 82 (2003) 1127–1137

system, coal particles sweep around the furnace volume and

have long residence times. In the primary combustion zone,
coal burns generally in conditions of low excess-air and
reduced firing rates. To consume the CO and unburnt
hydrocarbons, over-fire air is introduced in the burnout
zone.
In each corner, different burners are finely staged in
vertical direction, so that the precise combustion history,
including stoichiometry history and temperature history,
might be controlled. Staging the combustion air produces
fuel-rich/fuel-lean sequencing favorable for the conversion
of fuel bound N to N2; staging the coal injections controls
the NOx emission in the way that the NOx formed earlier in
the flame can be reduced by its reactions with hydrocarbon
radicals, so the NOx emission is fairly low, as shown in Refs.
[13,14].
Fig. 3 shows a simplified list of burners used for the
simulation, where SA represents different secondary air
inlets, PA denotes primary air inlets, ‘leaking air’ is a
controlled leakage air for some closed nozzles around the
exit of operating burners, and OFA indicates over-fire air
inlets. The air velocity is about 25 m/s at PA inlets, about
50 m/s at SA (including OFA) inlets at full-load operation.
The controlled leaking air velocity is about 5 m/s.
To model the discrete coal particles coupling to the
continuous gas flow, 24 coal injections are defined at the 24
PA inlets. Each of them is divided into 90 particle streams
according to nine different locations and 10 groups of initial
particle size, whose mass fractions obey to the Rosin –
Rammler distribution, with the mean diameter of 57 mm.
The detailed size distribution is shown in Table 2.
Beyond the combustion section, i.e. above the furnace
exit plane P1, there are three groups of panel super-heaters
along the main flow direction (i.e. in Y-direction), denoted in
Fig. 1 as SH1, SH2 and SH3, respectively. SH1 and SH2
have five plates each in X-direction, and are modeled as
constant-temperature (783 K) double-sided walls. SH3 has
Fig. 1. Simplified volume of the 609 MW boiler.
29 plates in X-direction, and is considered as a porous
medium whose inertial resistance coefficients in both Y- and
represent some lines where data are reproduced, in which Z-directions are set to 0.35 according to the actually
L0 – L3 are located on the furnace center-section in measured pressure drop; while that in X-direction is set to
Y-direction. All these marks are used frequently below. 50 to limit the anisotropy of these coefficients to two orders
Table 1 shows the analysis data of the coal fired. of magnitudes.
The primary air (including pulverized coal) and second- In the crossover pass, there arrange also three groups of
ary air are injected from burners located in the four corners panel super- and/or re-heaters in the main flow direction,
into the furnace at set angles to form imaginary circles in the denoted as HE1 – 3 in Fig. 1. All are modeled as porous
furnace center, as shown in Fig. 2. In such a concentric firing media, whose inertial resistance coefficients in both

Table 1
Analysis data of the coal fired

Proximate analysis (%, dry) Ultimate analysis (%, DAF) Lower heating value (MJ/kg, DAF)

Volatile Fixed carbon Ash C H O N S

35.1 47.41 17.49 85.01 4.79 7.18 1.70 1.32 33.066

 C. Yin et al. / Fuel 82 (2003) 1127–1137 1129

Y-direction and Z-direction are set to 0.72, 0.63 and 0.45,

respectively, based on the actually measured pressure
losses. Their resistance coefficients in X-direction are set
to a same value of 50. In the rear pass included, the group of
heaters is also modeled as porous zone, with the inertial
resistance set to 2, same in all the three directions according
to the actual operation. Radiation heat transfer is taken into
account for the porous mediums; while convective heat
transfer is not considered. Tests with assumed convective
heat transfer coefficients indicate that the inclusion of
convective heat transfer does not have an important effect
on the gas temperature deviation, although it affects the gas
temperature level nearby to a little extent.

3. Simulation methods

3.1. Mesh of the calculation domain

Fig. 2. Concentric firing system in the furnace. The calculation domain is shown in Fig. 1. Since the flow
is still quite unmixed at the furnace exit, and the gas
temperature deviation in upper part of the boiler is the
emphasis of this study, so the upper furnace, the crossover
pass and part of the rear pass are included in the simulations.
Special attention has been paid to grids and grid design to
reduce numerical error as much as possible [15]. A
structured mesh is created with boundary-fitted coordinates,
in which the grid lines are approximately along the swirl
flow direction in furnace, as shown in Fig. 2. Fig. 4 also
shows the calculated flow field at a plane in the furnace,
from which one can see that the mesh resolves the major
flow property in the furnace. The flow in the crossover pass
and the rear pass is relatively simple and therefore, the
commonly used quadrilateral grid is kept. There are totally
454,776 hexahedral cells, with 42 £ 37 £ 234 ¼ 363,636
cells in the furnace. The grid is fine enough to give grid-
independent solutions [11]. First-order spatial discretization
scheme is used for all equations. Validating the simulations
by some key global design parameters of the boiler and
available site operational records assures reliable and
acceptable simulation results, as well as the residuals
control.

3.2. Overall models

In the study, standard k – 1 model is used as closure of

turbulent Reynolds equations; SIMPLE method is used in
flow field prediction.
Lagrangian particle trajectory is traced. To take the
effects of turbulence on the particle trajectories into account,
stochastic tracking model are used in which three trajectory
calculations are performed to include turbulent velocity
fluctuations into the particle force balance. So totally
90 £ 24 £ 3 ¼ 6480 coal particle streams are tracked in
the calculations. During traveling through gas and interact-
Fig. 3. Burners at each corner. ing with gas, coal particles devolatilize and undergo char
1130 C. Yin et al. / Fuel 82 (2003) 1127–1137

Table 2
Size distribution of pulverized coal

Diameter (mm) 6 20 34 48 62 76 90 104 118 132

Mass fraction (%) 22.33 19.00 14.88 11.62 9.08 7.11 5.55 4.36 3.40 2.67

combustion, creating a source of fuel for reaction in gas the upper furnace, in the crossover pass and in the rear pass
phase. The two-competing-rates model is used to yield an are taken into account with the method described in Section
expression for the devolatilization. The heterogenerous 2. To investigate the effect of the panel super-heaters on the
combustion of char is calculated using kinetics/diffusion- gas temperature deviation, a second case, i.e. case 2, is also
limited model. The interaction between the gas and particle conducted, in which all platen heaters in the upper furnace,
phases is modeled by altering iterations between the flow in the crossover pass and in the rear pass are ignored. In this
calculation and the particle trajectory calculation, with 20 paper, case 1 is the default case for results presentation.
flow iterations per discrete phase iteration.
Species and chemical reactions are modeled using the
mixture-fraction/probability density function (PDF) 4. Results and discussions
approach and the full equilibrium chemistry, where the
turbulence-chemistry interaction is modeled using a double- 4.1. Calculation validation with boiler design values
delta PDF.
The discrete ordinates (DO) model is used to simulate The combustion calculation begins by solving the gas
radiation heat transfer, thanks to its ability to take into flow field equations assuming that the particles are absent.
account the particulate effects and localized heat sources, This simulation is well converged after more than 3000
the two important aspects of coal combustion simulation. iterations, as shown in Ref. [14], from which, full
All the models used in this study are general models, combustion simulations start.
widely used and well documented, and can be found in Refs. In the combustion calculations, the furnace exit par-
[2,16] for details. They are proven efficient to capture the ameters, such as gas temperature and O2 concentration on
complex phenomena in such a large-scale tangentially fired plane P1, are monitored, since they are very important for
boiler. The calculations in this paper just follow in the wake preserving the boiler operation and performance including
of the earlier studies [13,14]. emission, reliability and safety. For example, if furnace exit
gas temperature deviates from the design value, some
3.3. Calculation cases undesired condition might occur, such as, potential of
crossover pass tubes overheating, and increased slagging/
The basic case, i.e. case 1, is conducted on the basis of corrosion of water walls and super-heater tubes. The furnace
full-load operation. In this case, all the platen heaters in exit parameters are also used to validate the combustion
calculations. Table 3 shows the comparison of the
calculation (case 1) with some key design values, including
the furnace exit gas temperature and O2 mass fraction. The
designed furnace exit gas temperature and O2 mass fraction
are about 1600 K and 3.5%, respectively; and the total heat
transfer in the furnace is around 609 MW. The simulation
results in case 1 agree very well with the design values.
Here, it should be stated that the emissivity of the furnace
wall is set to 0.7 in this study for the radiation heat transfer
simulation, since the average temperature on plane P1 is
quite dependent on this parameter.

Table 3
Validation of combustion simulation with key global design values

Item In boiler design Calculated results

Average temperature on section ,1600 1602

P1 (K)
Average O2 mass fraction on section ,3.5 3.4
P1 (%)
Total heat transfer inside ,609 614
the furnace (MW)
Fig. 4. Velocity vectors at a plane in burnout zone.
 C. Yin et al. / Fuel 82 (2003) 1127–1137 1131

4.2. Gas temperature deviation characteristics

in this 609 MW boiler

Xu et al. gave the measured gas temperature deviation

between the two side-walls in a 600 MW boiler [12],
roughly corresponding to location L1 in this boiler. The
measurements were carried out twice to make sure the data
are reliable. Since the concentric firing system is in
clockwise direction in that 600 MW furnace, while it is in
counter-clockwise direction in this 609 MW furnace, the
indication of furnace side-walls is inversed here to make it
feasible to compare their measurements with this study. The
left and right side-walls used in this paper are shown in
Fig. 2, and a positive deviation indicates a higher value in
the left.
Fan et al. also gave some measured gas temperature
deviation in a large-scale tangentially fired boiler of similar
capacity [9], roughly corresponding to location L4 in this
boiler. It was concluded that the temperature deviation in
that boiler is not very serious because the boiler has just
been rebuilt.
Fig. 5 shows the gas temperature deviations on furnace
centerlines of different heights: L0 – L3 in this boiler. The Fig. 6. Gas temperature distribution on plane P3.
temperature deviation is not serious and is less than in the
reported cases. The gas temperature deviation increases
with the furnace height, and the maximum temperature the crossover pass. As shown in Refs. [13,14], the
deviation in this boiler, about 165 K, takes place in the temperature deviations on planes V2 and V3, where tube
location about 3 m away from furnace side-walls in zones explosion accidents might most probably take place due to
around line L3. It is believed that gas temperature deviation the higher steam parameters inside the tubes, are fairly
is inherent feature of tangentially fired boilers: typically small, less than 60 K: a little higher in the left. So, the focus
100 –150 K in 200 MW boilers, 150– 200 K in 300 MW below will be put on the temperature deviations in the upper
boiler, and 200– 250 K in 600 MW boilers. From this point, furnace.
one can also conclude the temperature deviation in this
609 MW boiler is acceptable. 4.3. Effect of uneven gas velocity distribution on gas
Fig. 6 shows the gas temperatures on plane P3, the temperature deviation
middle plane in the crossover pass in Z-direction. From it,
one can see the gas temperature deviations in the crossover Residual airflow swirling at furnace exit is a typical and
pass are less than those in the upper furnace. In fact, the gas inherent feature of tangentially fired boilers, and is believed
temperatures look symmetry between the two side-walls in to be one important cause for the temperature deviation. The
residual swirling leads to an uneven gas velocity distri-
bution between the two side-walls. In the upper furnace and
the crossover pass, the y-directional velocity, vy, is the
dominant component; and the next is the z-directional
velocity, vz; the magnitude and effect of the x-directional
velocity, vx, are found very small and can be ignored, except
in zones near the front-wall. To understand better how the
gas velocity is distributed and how the uneven velocity
distribution affects the temperature deviation, Fig. 7 shows
the velocity vectors on different planes in the upper furnace.
The gas flow in the platen super-heaters zone in the upper
furnace is much more complex than those inside the
combustion section.
As shown in Fig. 7, in the zones just beyond the
Fig. 5. Gas temperature deviation on centerlines of different heights: L0, L1, furnace exit, the gas velocity has a negative v y
L2 and L3. component in the left and a positive vy in the right
1132 C. Yin et al. / Fuel 82 (2003) 1127–1137

Fig. 7. Gas velocity vectors on planes in the upper furnace: P1, P2, P3 and P4.

because of the residual airflow swirling. With the height the upper furnace and in the crossover pass: low in the left
increases, the gas in the left gradually reverses its and high in the right.
direction under effect of furnace walls and super-heaters From Fig. 7, large vx components from the left to the
panels, and flows forward directly into the crossover right are found in the gaps between the front-wall and the
pass; while the velocity component, vy, in the right has a first group of platen super-heaters (SH1) in all planes of
slight decrease because less gas goes from the left into different heights in the upper furnace. This portion of gas is
the right and then into the crossover pass. from two sources. Part of the gas comes directly from the
Fig. 7 also exhibits a typical uneven gas velocity in the near front-wall zone in the combustion section, which goes
upper furnace: low in the left and high in the right. upward from there at a certain z-velocity into the
Combined with the vy component distribution in plane P3, corresponding zone in the upper furnace; another part is
as shown in Fig. 8, the formation mechanism of the uneven from the left in the upper furnace, which goes firstly
gas velocity in the upper furnace may be explained in Fig. 9, backward toward the front-wall and then into this zone.
where M1 denotes the momentum of the residual airflow Both of them are caused by the residual airflow swirling at
swirling in upper furnace; M3 is the momentum introduced the furnace exit, which is probably one of the most
by induction fan, which can be supposed as constant here important causes for the gas temperature deviation in the
along the width of crossover pass; and M2 represents the upper furnace: high temperature in the left and low
composite momentum of gas in crossover pass. That is temperature in the right, and this temperature deviation
the mainstream of the uneven gas velocity distribution in increases with the furnace height.
 C. Yin et al. / Fuel 82 (2003) 1127–1137 1133

Fig. 8. vy component on P3.

Fig. 9. Formation of gas flow deviation at plane P3.

At a relatively low level in the upper furnace, for

example on P2, the hot flue gas, leaving from the
With further increase of section level, for example on
combustion section, enters the upper furnace with upward
P4, the gas flow of negative vy component almost
velocity components. Because of the residual airflow
disappears. The upward gas flow, coming from the lower
swirling at the furnace exit, all the gas in the left goes levels, attenuates its vz component. Most of the gas in the
firstly in negative-direction; then turns into the gap left goes forward directly into the crossover pass with an
between the front-wall and SH1; together with the gas in increased vy component; while small portion of the gas in
the gap directly from the combustion section, it flows the left still enters the gap between the front-wall and the
toward the right with a high speed; cooled by the front-wall SH1 after resting in corner A, together with the gas
(the inner media temperature is averaged at about 603 K) existing directly from the lower levels. Same as above, this
and the SH1 (the average inner media temperature is about portion of gas firstly cooled by the front-wall and goes
718 K), the gas finally enters the panels channel in the from the right side into the crossover pass together with the
right, mixes with the existing gas there coming directly gas existing in the right.
from the combustion section, and goes into the crossover Since the same flow tendency is kept all along the height,
pass. Since the gas in the left is further cooled by the front- the gas temperature deviation will be accumulated and thus
wall and some SH1 panels before arriving in the right, a increase with the furnace height, as shown in Fig. 5.
temperature deviation is therefore observed: high in the left According to the above analysis, reducing the residual
and low in the right. swirling at the furnace exit will be an efficient way to
With increase of section level, for example on P3, the alleviate the temperature deviation in the upper
negative vy component in the left is attenuated under the furnace. Reasonable design of the concentric firing system
effect of the furnace walls and SH1 panels. Small portion of in the furnace and the distance between the upmost burners
the gas in the left might go forward directly into the to the furnace exit, and counter-offsetting part of the air-
crossover pass; while most of the gas still enters the gap injections can all affect the residual swirling at the furnace
between the front-wall and the SH1 with a high speed, exit. Except these measures, two new methods are put
passes through the panels in the right and then goes into the forward in this paper and might be helpful to alleviate the
crossover pass together with the gas there. deviation, based on the mechanism how the residual
1134 C. Yin et al. / Fuel 82 (2003) 1127–1137

Fig. 10. Two new methods to alleviate the gas temperature deviation in upper furnace.

swirling affects the temperature deviation in the upper the residual airflow swirling. This effect can be observed
furnace: the existence of high-speed gas flow from one side- from Fig. 11, in which the calculated deviations on lines of
wall to the opposite side-wall in the gap between the front- different heights in the two cases are shown.
wall and the platen super-heaters. The first is, a new nose The platen super-heaters attenuate the residual airflow
might be added on the front-wall near the furnace exit (P1 in swirling at furnace exit to some extent and reduce the
Fig. 1). In this way, the gas flow along the relatively cool velocity deviation between the two side-walls, as seen in
front-wall in the gap will be reduced under the directing of Fig. 12. In addition, super-heater panels can also attenuate
the nose. The second is, the super-heater panels can be re- the gas velocity imbalance in upper furnace and improve the
arranged to distribute uniformly the gas between two side- flow distribution.
walls. Taking this boiler as an example, they can be
described in Fig. 10. 4.5. Effect of particle trajectories and their combustion
histories on gas temperature deviation
4.4. Effect of platen super-heaters on gas
temperature deviation In tangentially fired boilers, the particles sweep around
the furnace volume, and it provides sufficient residence time
As stated in Ref. [12], the platen super-heaters in the for the very fine particles. During particle combustion, all of
upper furnace are another important cause for the gas the fuel species except char are consumed within 1 or 2 m of
temperature deviation. But, how the super-heater panels the burner throats, while char burns at a slower rate, and
affect the temperature deviation needs to be further studied. normally is not consumed until it reaches the furnace center.
From the above cause analysis of the temperature At the furnace exit cross-section (i.e. plane P1), the gas
deviation, it may be concluded that super-heater panels in temperature is fairly uniform between the two side-walls
the upper furnace and in the crossover pass can reduce the and the temperature deviation between the two side-walls is
temperature deviation to some extent since they attenuate very small, which was confirmed in Refs. [13,14] and can
 C. Yin et al. / Fuel 82 (2003) 1127–1137 1135

Fig. 11. Gas temperature deviation on line L0, L2 and L3 in different cases.

also be seen from the temperature plot on lines L0 – L3 in section (i.e. above the furnace exit), the char concentration
Fig. 13, thanks to the reasonable design of the concentric is very low, with the calculated mass fraction less than
firing system in the furnace, and the distance between the 2 £ 10210, and thus the effect of char combustion in the
upmost burner and the furnace exit, and thus the pretty long upper furnace is very small, which can also be confirmed by
period of intensive mixing in the furnace. Therefore, it may the very low mass-exchange between two phases. So char
be concluded that one can ignore the influence of the combustion in the upper furnace is very limited and may be
particle trajectories and their combustion histories inside ignored, which can also be reflected from the monotonous
the combustion section at all when studying the formation of decrease of gas temperature with increase of section height
the gas temperature deviation in the upper furnace. (shown in Fig. 13).
Although char combustion might still exist somewhere Due to the quite complex flow in the upper furnace, the
in the upper furnace, the zones with char combustion are particle concentration and enthalpy change are here exam-
very small and the combustion rate is also very low: about ined instead of particles trajectories to study how the particles
two orders of magnitudes lower than those in the affect the gas temperature deviation in the upper furnace. The
combustion section. Moreover, beyond the combustion particle concentration is found a little higher in the right,

Fig. 12. vy distribution along the furnace width on P1.

1136 C. Yin et al. / Fuel 82 (2003) 1127–1137

Fig. 13. Gas temperatures on some lines in upper furnace.

Fig. 14. Total enthalpy exchange from particles to gas in the upper furnace.

because gas velocity is a little higher in the right and particles the total mass flow rate in the boiler is about 660 kg/s. The
in upper furnace are very tiny. But this deviation is not accumulated enthalpy-exchange deviation itself in the upper
significant. Fig. 14 shows the exchange of total enthalpy from furnace is estimated to cause a low gas temperature deviation
the particles to the gas in the upper furnace. In fact, it is between the two side-walls, less than 15 K.
almost same as the sensible enthalpy exchange since the char So it may be concluded that the particle trajectories and
combustion is negligible. From the distribution of the combustion histories both inside and beyond the combustion
enthalpy exchange, one might be able to conclude that section do not have a significant effect on the temperature
the particles trajectories in the upper furnace will not cause a deviation in the upper furnace.
big gas temperature deviation, since both the transferred
enthalpy values and their deviations between the two side-
walls are fairly low, especially with increase of the furnace 5. Conclusions
height. If ignore all other factors, a 100 W deviation in the
enthalpy transferred from the particles to the gas will have Simulations are conducted to study the gas temperature
little effect on the gas temperature deviation since the specific deviations in a large-scale tangentially fired boiler,
heat of the gas is about 1300 J/(kg K) in the upper furnace and especially in the upper furnace. Good agreement between
 C. Yin et al. / Fuel 82 (2003) 1127–1137 1137

the simulation results and key design values of the boiler proposed to reduce gas temperature deviation in
and available site operation records indicates that the upper furnace in tangentially fired boilers, with special
calculations are reliable for such an investigation. Generally aims at reducing hot gas flow going from one side-wall
speaking, the gas temperature deviation in this 609 MW to another along the cool front-wall and distributing
boiler is not serious, thanks to reasonable boiler design. this gas flow uniformly between the two side-walls.
Some special conclusions can be drawn as follows.

(1) In a counter-clockwise concentric firing boiler, the gas Acknowledgements

temperatures in upper furnace and in crossover pass
feature such a deviation: high in the left and low in the The first author gratefully acknowledges Mr E. Perez in
right (see Fig. 2 for left and right indication). This ALSTOM Power Boilers, France, and Prof. Dr B. Baudoin
deviation is larger in upper furnace than in crossover in École des Mines de Douai, France, for their discussions.
pass.
(2) Residual airflow swirling at furnace exit is one
important cause for the gas temperature deviation. References
The residual swirling induces a high-speed gas flow
along the relatively cool front-wall, and thus causes a [1] Niska SA. Coal combustion modelling. IEA Coal Res 1996; IEAPER/
temperature deviation between the two side-walls. 31.
[2] Eaton AM, Smoot LD, Hill SC, Eatough CN. Prog Energy Combust
(3) The gas temperature deviations in upper furnace
Sci 1999;25:387.
increase with the height. The gas flow is much more [3] Boyd RK, Kent JH. The 21st International Symposium on Combus-
complex in upper furnace than that in combustion tion. Pittsburgh, PA: The Combustion Institute; 1986. p. 265.
section: the gas velocity deviation in upper furnace [4] Bonin MP, McQuay M. Fuel 1996;75:195.
does not necessarily decrease with the furnace height; [5] Fiveland WA, Latham CE. Combust Sci Technol 1993;93:53.
[6] Liu X, Xu X, Fan H. Proc CSEE 2001;21:80.
the gas temperature will also not necessarily be high in
[7] Lookwood FC, Mahmud T, Yehia MA. Fuel 1998;77:1329.
zones where gas velocity is high. [8] Stanmore BR, Visona SP. Fuel Process Technol 2000;64:25.
(4) Platen super-heaters in upper furnace are also one [9] Fan J, Qian L, Ma Y, Sun P, Cen K. Chem Engng J 2001;81:261.
cause for the temperature deviation. They alleviate, [10] Fan J, Sun P, Zha X, Cen K. Energy Fuels 1999;13:1051.
while not aggravate, the deviation since they can [11] Fan J, Sun P, Zheng Y, Ma Y, Cen K. Fuel 1999;78:1387.
[12] Xu M, Yuan J, Ding S, Cao H. Comput Methods Appl Mech Engng
attenuate the residual airflow swirling and improve the
1998;155:369.
gas velocity distribution; and this effect is limited. [13] Yin C. Numerical simulation of the flow, combustion, heat-transfer
(5) Particle trajectories and their combustion histories and emissions from a 609 MW tangentially fired pulverized-coal
inside and beyond the combustion section can affect boiler. Post-Doctoral Research Report, Département Energétique
the gas temperature deviation in upper furnace, but Industrielle, École des Mines de Douai, France; 2000.
[14] Yin C, Caillat S, Harion J, Baudoin B, Perez E. Fuel 2002;81:997.
they are just a minor factor and their effect is not
[15] Casey M, Wintergerste T. ERCOFTAC special interest group on
important. quality and trust in industrial CFD—best practice guidelines; 2000.
(6) Two new methods, a nose on the front-wall and re- [16] Smoot LD, Smith PJ. Coal combustion and gasification. New York:
arranged super-heater panels, are unprecedentedly Plenum Press; 1985.