Fuel 107 (2013) 688–698

Contents lists available at SciVerse ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Numerical simulation of brown coal combustion in a 550 MW

tangentially-fired furnace under different operating conditions
Audai Hussein Al-Abbas a, Jamal Naser b,⇑, Emad Kamil Hussein a
a
Foundation of Technical Education, Al-Musaib Technical College, Babylon, Iraq
b
Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, Hawthorn, Victoria 3122, Australia

h i g h l i g h t s

" CFD modeling study was performed for the combustion of the brown coal in a large-scale tangentially-fired furnace.
" Performance of the boiler under ten different operating conditions was investigated.
" The temperature distributions were better when the turned off burners are set in the opposite direction.
" The results showed improvements on the combustion characteristics in comparison with the standard operating case.

a r t i c l e i n f o a b s t r a c t

Article history: In the present paper, a computational fluid dynamics (CFD) modeling study was performed for the com-
Received 9 September 2012 bustion of the brown coal in a large-scale tangentially-fired furnace (550 MW) under different operating
Received in revised form 13 November 2012 conditions. The AVL Fire CFD code has been used to model the combustion processes. The mathematical
Accepted 14 November 2012
models of coal combustion with the appropriate kinetic parameters were written and incorporated to the
Available online 8 December 2012
code as user defined functions. These models consist of pulverised coal (PC) devolatilization, char burn-
out, and heat and mass transfer. The simulation of the PC combustion was carried out using multi-step
Keywords:
reaction chemistry mechanisms. The level of confidence of this numerical model was based on the pre-
Coal combustion
Victorian brown coal
vious validations of the lignite combustion in a lab-scale furnace, as well as the validation parameters of
Emissions the present furnace at the standard existing conditions in terms of temperature values and species con-
CFD centrations. Performance of the boiler under ten different operating conditions was investigated. The
strategy of operation schemes for the first six combustion scenarios were based on the change of the
out-of-service (turned off) burners under full load operation, while the rest cases were carried out at
20% lower and 20% higher loads than the standard operating conditions. The validated model was used
to perform the following investigation parameters: furnace gas temperatures, species concentrations
(O2, CO and CO2), velocity distributions, and char consumption. The predictions demonstrated that there
are good temperature distributions in the furnace when the turned off burners are set in the opposite
direction under full load operation. For higher aerodynamic effect, the numerical results showed
improvements on the combustion characteristics in terms of species concentrations and char burnout
rates in comparison with the standard operating case. The findings of this study provide good information
to optimize the operations of the utility tangentially coal-fired boiler with less emission.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction (GHGs) emissions and a high moisture content which is equal to

about 60–70 wt% [6]. However, In order to obey the new environ-
The brown coal combustion in the tangentially-fired power mental and political legislation against global warming, it is neces-
plants is the main source of electricity production in Victoria/Aus- sary to find out a cost-effective solution to cut pollution.
tralia [1–4]. Although the abundance, low-cost, high reactivity, and The understanding of the brown coal reactivity and behavior
low sulfur content of the brown coal [5] the major disadvantages of under different operating conditions is required to design clean
this source of energy are its contribution to the greenhouse gases and efficient brown coal combustion systems. Computational fluid
dynamics (CFD) modeling studies can comprehensively provide a
wide range of information for optimizing and improving the com-
⇑ Corresponding author. Tel.: +61 3 9214 8655. bustion characteristics and boiler performance that can reduce the
E-mail address: [email protected] (J. Naser). cost of time-consuming experimental tests. The flame structure,

0016-2361/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2012.11.054
 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698 689

temperatures distributions, chemical species concentrations, nitric combustion cases were dependent on the scenarios of change in
oxides, radiative heat flux, etc., under different firing conditions the mass flow rates and distribution ratios for the PC, mill gases
can be carried out using the appropriate CFD code [7–10]. The work and air (i.e. 20% lower and 20% higher than the standard combus-
programme covered by this paper achieves the development and tion conditions). In addition, the operation schemes of turned off
validation of the computational tool applied to the combustion (out-of-service) burners were also examined under full load oper-
system which indicates the practical relevance and the long term ation. The species concentrations (O2, CO2, and CO), furnace gas
viability for such a tool. These developments allowed the complete temperature distributions, velocity fields, and char consumption
simulation of commercial large-scale power plants as an effective obtained for all combustion cases were compared. The combustion
alternative to expensive experimental tests. characteristics and boiler performance are considerably depended
In general, after the devolatilization process of the coal particles upon the turbulent mixing conditions, residence times of combus-
finishes, the char combustion subsequently starts in the firing sys- tibles, and aerodynamic effects used in this numerical study.
tem. This combustion process has been considered the dominant
factor constraining several reaction parameters such as the total
burnout time, unburned carbon level, and radiation from burning 2. Geometry and operating conditions of the boiler
char particles [11–16]. A better understanding of the effects of
these parameters on the combustion characteristics and boiler heat Loy Yang A power plant located in the state of Victoria in
transfer under different operating conditions enables engineers to Australia was selected to use in the present numerical investiga-
optimize the applications of both the existing and new coal power tions. The present tangentially-fired furnace is composed of eight
plants. Ash deposition and slag formation from burning pulverized mill-duct systems, which include six separate burners for each duct
coal particles are significant parameters in the thermal boiler de- system. On each side of the furnace there are two firing groups,
sign, particularly on the heat transfer surfaces such as radiation including three inert burners in the upper location and three main
zone (water–wall furnace) and convection zone (superheater, burners in the lower location of the furnace zone, as well as many
reheater, and economizer) [8]. Therefore, it is necessary to under- secondary air ducts for each group. The top–bottom sequence
stand how these parameters vary when the rates of pulverised coal arrangement of each firing group on the furnace wall can be found
(PC) and feed oxidizer gases are changed in the furnace inlets com- in the previous numerical studies [19,20]. In the convection zone,
pared to the existing operating conditions. However, the full there are eight sources of heat sink located above the hot gas off
knowledge of the effects of these intrinsic parameters on the com- take (HGOT) pipes. The arrangement of the convection tubes in
bustion characteristics and power plant efficiency is still not com- the boiler are economiser, reheater 1a, superheater 2, reheater 1b,
pletely available. superheater 3, reheater 2, superheater 4, superheater 1, and their
Tian et al. [17] developed a CFD model to investigate the coal heat absorption values are 100.2, 121.98, 151.74, 121.98, 192.12,
combustion in a 375 MW tangentially-fired utility boiler, using 98.48, 52.27, 68.44 MW, respectively. The heat absorption value
ANSYS CFX 12.0 code. Due to the CFD code limitations regarding of water wall is 458.75 MW, and its location is beneath HGOT pipes.
the multi-step reaction of hydrocarbon released, the authors uti- The effect of these heat sink sources are considered and taken into
lized carbon monoxide reaction with oxygen to improve the model account of the calculation as a cell selection in the computational
predictions. After validating the CFD model with the measured domain. Around 430 kg/s of steam flow is produced in this power
data obtained from the Yallourn power plant, then they used the unit under full load operation at 16.8 MPa and 540 °C.
validated model to examine the flame shape, temperature distribu- The mesh generation with the geometric description of the CFD
tion, and wall incident heat flux. The numerical investigations were model used for the boiler is shown in Fig. 1a. The geometric dimen-
done under different combustion air distribution scenarios and at sions of the furnace are taken from the drawings of the power
different out-of-service firing groups. The predictions showed that plant, and CAD tool is employed to construct the 3D geometric
there is a significant difference in the combustion characteristics in model. The dimensions of the simulated furnace were 98.84 m
the furnace, and these differences completely dependent upon (height), 17.82 m (width), and 17.82 m (depth). Under the standard
which burner groups are turned off under full load operation. operating conditions, the mills 1, 2, 5, 6, and 8 are in service, while
A validated 3D numerical model of the lignite combustion in a the rest mills (3, 4, and 7) are out-of-service, which presented for
300 MW large-scale utility boiler under the standard firing condi- the combustion case 1, as illustrated in Table 1. Table 1 presents
tions was used by Karampinis et al. [18] to perform a numerical the operation scheme of turned off (out-of-service) burners under
investigation for lignite coal and biomass particles at several co-fir- full load operation (cases 1–6) and 20% lower (cases 7 and 8) and
ing scenarios. The numerical findings showed that the burnout 20% higher (cases 9 and 10) than the standard operating condi-
rates of lignite char coal decreased in the hopper under co-firing tions. Around 81.3 kg/s of PC and 422.8 kg/s of gas mixture are
conditions. While for biomass fuel, the small diameter of particles passed through both the inert and main burners of the furnace at
resulted in an increase in the char burnout for all burner levels different flow distribution ratios under the standard operating
used. The authors concluded on two main concepts for the poten- conditions (as specified in cases 1–6). More detailed information
tial operation schemes. The first was to mix the biomass particles about the validated results of the CFD model against the power
with the lignite stream to ensure a better biomass size reduction. plant data and boundary conditions used in the standard
And the second concept was dependent on the fact that the injec- combustion conditions were given in the recent published paper
tion system of the biomass particles should be placed between the of Al-Abbas et al. [20].
lower and upper main burner to optimize the char burnout in the Table 2 shows the mass flow rates (kg/s) of air for the standard
boiler. firing case and investigated combustion cases at each secondary air
Up to date, in the field of numerical simulation on the commer- duct. The overall number of vapor and PC burners was 48, while 18
cial large-scale facility firing unit, there has been unfortunately lit- of the total burners were practically out of service for all combus-
tle research work conducted on the brown coal combustion. tion cases examined. Regarding the burner configuration, each bur-
Therefore, the objective of this study is to simulate the brown coal ner ports 1, 3, 5, and 7 were inclined by 24° with the perpendicular
combustion in a large-scale tangentially-fired furnace under line to the furnace face, while each the remaining burner ports 2, 4,
several operating conditions. A computational fluid dynamics 6, and 8 were inclined by 30°. In the most types of the tangentially-
(CFD) code, AVL Fire version 2009.3, was used to model and fired furnaces/boilers, this configuration of the burner set up was
analyze ten different combustion environments. The investigated typically used in order to improve flame stability inside the
690 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698

3. Mathematical models and numerical description

Numerical modeling of the pulverized Victorian brown coal

under different combustion conditions were carried out using a
computational fluid dynamics (CFD), AVL Fire 2009.3 code [21].
The CFD code was used to solve the Eulerian partial differential
equations (PDEs) [22] for mass, momentum, enthalpy, a number
of species mass fractions, and turbulent fields in the gas phase
flow. The Discrete Droplet Method (DDM) [23] was employed
solving ordinary differential equations for the solid phase flow.
The subroutines required for the devolatilization and char burnout,
trajectories, convection and radiation heat transfer between parti-
cles and gases were written and incorporated into the CFD code as
sub-models of source terms of PDEs. The standard k–e model is
used modeling the turbulent flow calculation. This model has been
demonstrated a sufficient accuracy in the near-burner region of
previous modeling simulation of lignite coal combustion [24].
The mixing rate of species is controlled using the eddy-breakup
(EBU) turbulent combustion model in this coal reaction process.
The proximate and ultimate properties of the brown coal particle
are listed in Table 4. Regarding the heat transfer models, the
discrete transfer radiation method (DTRM) has been used because
of its ability for a better prediction in participating media,
especially in furnaces.
The chemistry reaction mechanisms for both the volatile mat-
ters and char coal are conducted. In this numerical study, methane
is considered as a devolatilized hydrocarbon fuel because there is
no essential differences can be distinguished between methane
and the gases produced from devolatilization processes of the coal
particles. However, the multi-step reaction mechanisms are car-
ried out in this numerical calculation. The main reactions of the
coal combustion model can be, however, expressed in three homo-
geneous and three heterogeneous chemical reactions, as follows:
The chemical equations for devolatilized methane burned with
Fig. 1a. CFD geometrical model of unit 1 at Loy Yang A power plant. oxygen, in a three-step reaction, are given in below:
CH4 þ O2 ! CO þ H2 þ H2 O þ heat ð1Þ
furnace, as schematically plotted in Fig. 1b. The mass flow rates of
both the PC and mill gases (recycled flue gases) through the main CO þ H2 O ¡ CO2 þ H2 ð2Þ
and inert burners for the standard, 20% lower, and 20% higher com-
bustion scenarios are presented in Table 3. O2 þ 2H2 ¡ 2H2 O ð3Þ

Table 1
The operation scheme of turned off (out-of-service) burners under full load operation (cases 1–6) and 20% lower (cases 7 and 8) and 20% higher (cases 9 and 10) than the standard
operating conditions.

Combustion case no. 1 2 3 4 5 6 7 8 9 10

Burners turned off 3, 4, and 7 3,4, and 6 3, 4, and 8 1, 3, and 4 3, 5, and 7 2, 4, and 6 3, 4, and 7 3, 5, and 7 3, 4, and 7 3, 5, and 7
Combustion scenarios Standard operating conditions Lower by 20% Higher by 20%

Table 2
The mass flow rates (kg/s) of air for the standard case and investigated combustion scenarios at each secondary air duct.

Secondary air duct Distribution ratio (%) Air standard 20% Lower 20% Higher
Mass flow (kg/s) Mass flow (kg/s) Mass flow (kg/s)
Upper inert 5.0 6.89 5.51 8.27
Upper intermediate inert 5.0 6.89 5.51 8.27
Lower intermediate inert 5.0 6.89 5.51 8.27
Lower inert 5.0 6.89 5.51 8.27
Upper main 20.0 27.55 22.04 33.07
Upper core 6.67 9.19 7.35 11.03
Intermediate main 20.0 27.55 22.04 33.07
Intermediate core 6.67 9.19 7.35 11.03
Lower main 20.0 27.55 22.04 33.07
Lower core 6.67 9.19 7.35 11.03
Total 137.78 110.21 165.38
 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698 691

Fig. 1b. The schematic representation of the burners’ configurations.

Table 3
Mass flow rates of both the PC and mill gases through the main and inert burners for the standard, 20% lower, and 20% higher combustion scenarios.

Burner duct Mass flow rates (kg/s) of PC for combustion cases Mass flow rates (kg/s) of mill gases for combustion cases
Standard 20% Lower 20% Higher Standard 20% Lower 20% Higher
Upper inert 0.93 0.74 1.11 18.71 14.95 22.42
Intermediate inert 0.54 0.43 0.64 18.22 14.54 21.82
Lower inert 1.57 1.22 1.83 18.19 14.54 21.82
Upper main 4.82 3.86 5.80 10.3 8.25 12.38
Intermediate main 2.78 2.24 3.35 9.34 7.51 11.26
Lower main 5.62 4.50 6.77 9.8 7.85 11.77
Total 16.26 13.01 19.51 84.56 67.65 101.47

While the chemical equations for char burned with oxygen, car- was employed solving the combination between the velocity and
bon dioxide, and water vapor are written as follows: pressure in the Navier–stokes equations. For the gas–solid two
phase flow, a Lagrangian/Eulerian approach was utilized. Around
Cchar þ 1=2O2 ! CO þ heat ð4Þ 50,000 particles were used in the present large-scale furnace
simulation. Further detailed information in regards to the mathe-
Cchar þ CO2 ! 2CO ð5Þ matical models and numerical description used in this study can
be found elsewhere [20,22,24,25].
Cchar þ H2 O ! CO þ H2 ð6Þ
The heats of combustion of the above-mentioned chemical 4. Results and discussion
equations for methane and residual char can be found elsewhere
[22]. This three-step chemical reaction scheme is used in this study 4.1. Temperature distribution
based on the good prediction results against the measured data of
the previous simulation studies of the coal combustion [20,22,25]. Fig. 2 presents the effects of different operating conditions on
Regarding the numerical description, the solutions of the math- the flame shape in the combustion zone at the 1650 K iso-surfaces
ematical equations of the heat transfer, coal combustion, and tur- for all combustion cases examined. The effects of different turned
bulence were conducted under transient mode. The numerical off burners operations under full load are presented in panels
results were run up to 48,000 time-steps, averaging the results (a–f), while in panels (g–j) the effects of mass flow rates and
over the final 8000 time-steps which were reached to the stable distribution ratios for the brown coal particles and mill gases are
quasi steady state. The convergence limitation for all variables investigated. That was done by using the validated CFD model
was attained with less than 104. The standard SIMPLE algorithm [20] under two different combustion scenarios: 20% lower (cases
692 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698

Table 4
The proximate and ultimate properties of the brown coal particle.

Proximate analysis (%wt, db) Ultimate analysis (%wt, db) Specific energy (MJ/ Moisture content (M.C.) (%wt, ar)
kg)
Fixed carbon Ash content Volatile content C H N S Minerals and inorgs O Gross dry Net wet 62.0
47.8 1.7 50.5 67.7 4.63 0.52 0.3 1.89 24.96 27.6 8.8

Fig. 2. Iso-surfaces of the combustion flames at 1650 K: (a) case 1, (b) case 2, (c) case 3, (d) case 4, (e) case 5, (f) case 6, (g) case 7, (h) case 8, ((i) case 9, (j) case 10.

7 and 8) and 20% higher (cases 9 and 10) than the standard com- turned off burners. A good distribution of the flame was basically
bustion case, as shown in Tables 1 and 2. This iso-surface temper- achieved in cases 3, 5 and 6 because of turning off the opposite
ature can clearly show the flame distributions in the furnace and burners. In panels (g) and (h), the flames are concentrated in the
determines, as a result, the boiler performance under different near-burner region due to the reduction in the mass flow rates of
combustion conditions. The operation scheme of the combustion combustibles. Whereas the combustion flames are uniformly dis-
cases used in this study was basically dependent on the inclined tributed in the burner and water wall regions for both cases 9
angles of each burner ports, as schematically plotted in Fig. 1b. and 10 due to the higher aerodynamic effects of the PC, mill gases,
However, this strategy of testing is used to improve the turbulent and air adopted in these two cases.
mixing of PC with the feed oxidizer gases in the tangentially-fired Fig. 3 presents the temperature distributions at the cross-sec-
furnace which essentially dependents on the location of the central tion cut (X–Y plane) of the lower intermediate main (LIM) burner
vortex in this type of furnaces. This improvement on the combus- for six combustion cases (cases 1–6 as specified in Table 1). In
tion conditions can bring several benefits to the combustion char- these combustion cases, the central vortexes of the flames are
acteristics and reduces the fouling and slagging problems on the highly affected by the turned off burners adopted in this study.
surfaces of the heat exchanges. Regarding the cases investigated In cases 1–4, there was movement toward the adjacent turned
under the strategy of turned off burners (cases 1–6 as specified off burners, this can be clearly seen in the next subsection of the
in Table 1), it can be seen that the flame temperature distribution velocity distribution. This may lead to a significant effect on the
of case 2 showed a clear tendency toward the out-of-service burn- temperature field symmetry and, as a result, on the heat transfer
ers (eastern water wall side in the furnace). This feature might be distribution in the furnace. In contrast, in cases 5 and 6, the central
caused due to the movement of the central vortex toward the vortexes are approximately concentrated in the central point of the
 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698 693

Fig. 3. Temperature distributions on the lower intermediate main burner at the cross-section cuts (X–Y plane) for six combustion cases: (a) case 1, (b) case 2, (c) case 3, (d)
case 4, (e) case 5, (f) case 6.

Fig. 4a. Temperature distributions for combustion cases (1, 7, and 9) along the Fig. 4b. Temperature distributions for combustion cases (5, 8, and 10) along the
central line of the furnace in the burners and water wall regions. central line of the furnace in the burners and water wall regions.

furnace. This can lead to a good turbulent mixing between the PC the temperature distribution along the Z-axis (X and Y = 0.0) of
and feed oxidizer gases and also helps to improve the stabilization the furnace in the regions of burners and water tube wall is shown
of the flame position. Therefore, it can be concluded that the pre- in Figs. 4a and 4b, compared with the standard combustion condi-
dicted results showed that the turned off burners in the opposite tions. In panel a of Figs. 4a and 4b, the numerical results presented
direction are better than those in the adjacent direction under full for combustion cases 1, 7, and 9, where the burner numbers 3, 4,
load operation. and 7 were out-of-service (see Table 1). While in panel b of the
The influence of changing the mass flow rates of pulverized coal same figure, the predictions were for combustion cases 5, 8, and
particles and feed oxidizer gases (recycled flue gases and air) on 10, where the burner numbers 3, 5, and 7 were out-of-service.
694 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698

Fig. 5. Gas velocity vector on the upper intermediate inert of the secondary air duct for six combustion cases: (a) case 1, (b) case 2, (c) case 3, (d) case 4, (e) case 5, (f) case 6.

The predictions showed that the effects of combustibles (fuel/ In these last two cases, the tangential distributions of the gas
gases) ratios, decreasing or increasing, have similar trends on velocity vectors (central vortexes) were approximately close to
temperature distributions with the standard operating conditions. the central zone of the furnace compared to the other combustion
In general, in both panels a and b, there were two maximum tem- cases. As a result, this good tangential circulation of the gases led to
perature values in the main burner and inert burner regions. This is improve the flame distribution, as seen in Fig. 2. In contrast, the
because of the higher amounts of coal particles and oxygen avail- central vortex in the case 2 was away from the central position
able in these regions of the furnace. Based on the fuel/gases ratio of the furnace and skews toward the furnace wall. The aerody-
adopted in this study, the combustion scenario (cases 9 and 10) namic effects, applying in cases 7 and 9, on the mean gas velocity
with 20% higher than the standard operating conditions showed along the centerline of the furnace in the regions of burners and
approximately the same temperature values in the burner region water wall (Z = 0.0–40.0 m) are showed in Fig. 6, compared to the
(Z = 0.0–15 m). But, there was a slight decrease in the temperature combustion case 1 that works under the standard operating condi-
profiles of the last combustion cases in the water tube wall region. tions. As reportedly mentioned in Table 3, the values of the mass
This can be explained due to the fact of the heat absorption by the flow rates in the inert burners are approximately double those in
water wall (sink source), as well as the additional cooling from the main burners in all combustion scenarios adopted in this study.
injecting extra cold air in the secondary air ducts. On the other
hand, in cases 7 and 8, the temperature profiles were lower than
the standard combustion conditions due to the reduction in the
aerodynamic flows in the combustion zone. It is interesting to note
that the average values of gas temperatures, in Figs. 4a and 4b,
were 1681.37, 1604.69, and 1647.37 K for combustion cases 1, 7,
and 9, respectively and 1731.95, 1667.2, and 1701.7 K for combus-
tion cases 5, 8, and 10, respectively. Compared to cases 7 and 8, the
average gas temperatures of combustion cases 9 and 10 were more
close to those of cases that work under the standard operating con-
ditions. For maintaining the heat transfer performance of the boiler
with less emission, cases 9 and 10 can achieve that purpose, as will
be shown in the subsections of species concentrations and char
consumption.

4.2. Velocity distribution

Fig. 5 presents the cross-section cuts of the gas velocity vectors

on the upper intermediate inert (UII) of the secondary air duct for
the combustion cases (cases 1–6 as specified in Table 1) that work
under the standard operating conditions. With reference to Fig. 3, Fig. 6. Mean velocity (m/s) along the centreline of the furnace in the region of the
cases 5 and 6 showed good flow field distributions in the furnace. burners for three combustion cases (1, 7, and 9).
 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698 695

Fig. 7. Oxygen mass fraction (kg/kg) along the hopper and burner regions of the furnace at the mid cut (X–Z plane) for three combustion cases: (a) case 1, (b) case 7, (c) case 9.

This gives a clear explanation why the mean gas velocity is high in 4.3. Species concentrations
the point Z = 14.0 m, which represents the level of inert burners in
the furnace, and it is low in the main burners region. However, in In Fig. 7, the distributions of O2 mass fraction (kg/kg) are pre-
Fig. 6, the maximum values of the mean gas velocity were 25.27, sented along the hopper and burner regions of the furnace at the
22.59, and 29.41 m/s for combustion cases 1, 7, and 9, respectively. mid cut (X–Z plane) for three combustion scenarios 1, 7, and 9. In
In the hot gas off takes (HGOTs) region, the reduction in the veloc- the zones of main and inert burners, there was a clear consumption
ity is evident for all combustion cases examined. in the oxygen concentrations for all cases investigated compared to

Fig. 8. Carbon dioxide mass fraction (kg/kg) along the height of the furnace at the mid cut (X–Z plane) for three combustion cases: (a) case 1, (b) case 7, (c) case 9.
696 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698

Fig. 9. Carbon monoxide mass fraction (kg/kg) along the height of the furnace at the mid cut (X–Z plane) for three combustion cases: (a) case 1, (b) case 7, (c) case 9.

the other regions in the furnace. This reduction in O2 was accom-

panied by high temperature values due to the intense combustion
reaction occurred in this region, as earlier shown in Fig. 4a. The
lower temperature profile observed in the combustion case 7
was connected with the higher O2 content noticed in Fig. 7, and
vice versa for the standard combustion case. Compared to cases 1
and 7, the higher O2 concentration was observed in the hopper
zone of the combustion case 9 because of the higher contents of
oxygen entering through the burners and secondary air ducts, i.e.
20% higher than the standard combustion conditions. Therefore,
it is important to mention here that these O2 concentrations are
strongly dependent on the aerodynamics and reaction conditions.
Fig. 8 shows the distributions of carbon dioxide (CO2) mass frac-
tion along the height of the furnace at the mid cut (X–Z plane) for
three combustion cases 1, 7, and 9. The difference in the CO2 con-
centrations between the standard (reference) combustion case and
two investigated cases are evident, and it is comparable in this
study. In Fig. 8, it can clearly see the reduction in the CO2 for both
combustion cases 7 and 9 compared to the conventional case. The
reasons for this decrease can be explained as follows: in case 7 the
load is decreased by reducing the PC, mill gases, and air flow rates
by 20% for each inlet port burner in comparison with the conven-
tional combustion conditions, and this can essentially increase the
residence time of coal particles in the combustion zone, and there-
by improve the coal reaction with oxidizers. This decrease of the
CO2 concentrations was also observed in the similar numerical
study of Belosevic et al. [26]. They used a comprehensive three-
dimensional mathematical model and CFD code to predict the
complex processes in two-phase turbulent reactive flows of pul-
verized coal particles within a large-scale utility boiler under dif-
ferent operating conditions. Similarly, Agraniotis et al. [27]
numerically investigated the co-combustion of brown coal in large
scale boilers under different operational conditions. In regards to
Fig. 10. (a) Mass fraction (MF) of carbon dioxide for combustion cases (2, 3, 5, 6, 9,
the reduction of CO2 observed in the combustion case 9, it can be and 10) along the central line of the furnace in the burner regions; (b) mass fraction
said that the reduction happened due to the higher O2 content of carbon monoxide for combustion cases (2, 3, 5, 6, 9, and 10) along the central line
available in the furnace which led to efficiently enhance the of the furnace in the burner regions.
 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698 697

reaction processes, as shown in Eqs. (1) and (4). However, the con- 4.4. Char consumption
centrations of carbon dioxide at the furnace exit were equal to
18.84, 16.65, and 16.19 (wt.%) for combustion cases 1, 7, and 9, In general, the numerical simulation of brown coal combustion
respectively. As mentioned previously, the combustion case 9 can in a tangentially-fired furnace is considered a complex process be-
provide a good opportunity to reduce emissions and maintain cause of the higher water content in this type of low-rank coal
the heat transfer performance of the boiler compared to the exist- (>60% water content in the raw coal). After the evaporation process
ing combustion case. In contrast to the lower CO2 concentrations from coal particles in the hot gas off take tubes (recirculation pro-
observed in cases 7 and 9, the latter two combustion cases showed cess), the thermal decomposition and the sequence burnout of the
a slight increase in the carbon monoxide (CO) concentrations in the released volatile matter and char coal are started in the mouth of
combustion zone relative to the standard firing case, as presented the furnace. As previously noted in the lignite coal combustion
in Fig. 9. This was likely due to the thermal dissociating mechanism within the lab-scale furnace and in the present furnace as well,
adopted in this study that led to an increase in the reactive pro- the reaction of the released hydrocarbon (volatile matter) with oxi-
cesses between the main (CO2 and H2O) and intermediate (H2 dizers was started and totally finished in vicinity to the burner re-
and CO) chemical species. In addition, the three-step reaction gion. And this resulted in an increase in the gas temperature.
mechanisms used in the heterogeneous reaction of char with the Therefore, the char consumption is the main interesting point that
oxygen, carbon dioxide, and water vapor (see Eqs. (4)–(6)) have a could be used to distinguish among the examined combustion
significant effect on the CO concentrations in the hot furnace zone. cases in this numerical investigation.
Fig. 10a and b presents the distributions of mass fraction of CO2 Fig. 11 shows the char content of brown coal particles (%) along
and CO respectively along the central line (X and Y = 0.0) of the fur- the height of the furnace for three combustion cases 1, 7, and 9. A
nace in the region of combustion zone (burners region) for differ- clear difference can be seen in the char burnout for these examined
ent combustion cases: 2, 3, 5, 6, 9, and 10. Both figures show combustion cases. This is basically due to the turbulent mixing
that there are two peak values of CO2 and CO concentrations for conditions, residence times of coal particles, and aerodynamic ef-
all combustion cases examined. The maximum values were firstly fects. Compared to the existing combustion case, cases 7 and 9
close to the main burners (PC burners), while the second max. val- showed a considerable increase in the char consumption. The rea-
ues were in the inert burner region. These two values resulted from son of this is that in the case 7 the higher residence time of coal
the availability of the higher O2 concentrations in these regions. In particles in the furnace zone gave a sufficient time to burn much
the case 6, the concentrations of CO2 and CO were somewhat lower char coal, whereas the higher aerodynamics and oxygen content
than those of the other combustion cases. As mentioned earlier, in the case 9 led to efficiently improved mixing and reaction con-
this improvement on the emission reduction might be due to the ditions. This is of a great importance to reduce the slagging and
good mixing condition adopted in this combustion case. fouling problems on the heat transfer surfaces. This interesting

Fig. 11. Char content of the pulverized coal particles (%) of three combustion cases: (a) case 1, (b) case 7, (c) case 9.
698 A.H. Al-Abbas et al. / Fuel 107 (2013) 688–698

result on the char burnout can lead to an increase in the combus- [4] Achim D, Naser J, Morsi YS, Pascoe S. Numerical investigation of full scale coal
combustion model of tangentially fired boiler with the effect of mill ducting.
tion efficiency under these new firing conditions.
Heat Mass Transf/Waerme Stoffuebertragung 2009:1–13.
[5] Li C-Z. Adv Sci. Victorian Brown Coal. Oxford: Elsevier; 2004.
[6] Tian ZF, Witt PJ, Schwarz MP, Yang W. Combustion of predried brown coal in a
5. Conclusion tangentially fired furnace under different operating conditions. Energy Fuels
2012;26(2):1044–53.
A 3D CFD model of a 550 MW tangentially-fired furnace at Loy [7] Habib MA, Ben-Mansour R, Abualhamayel HI. Thermal and emission
characteristics in a tangentially fired boiler model furnace. Int J Energy Res
Yang A power plant has been developed. The validated model was 2009. http://dx.doi.org/10.1002/er.1667.
used to investigate the combustion characteristics and boiler per- [8] Nikolopoulos N, Nikolopoulos A, Karampinis E, Grammelis P, Kakaras E.
formance under different firing conditions. The combustion sce- Numerical investigation of the oxy-fuel combustion in large scale boilers
adopting the ECO-Scrub technology. Fuel 2011;90(1):198–214.
narios of this work were based on the effect of the mass flow [9] Fang Q, Musa AAB, Wei Y, Luo Z, Zhou H. Numerical simulation of multifuel
rates and distribution ratios for the PC, mill gases, and air in the combustion in a 200 MW tangentially fired utility boiler. Energy Fuels
combustion zone compared to the reference scenario. Ten different 2012;26(1):313–23.
[10] Habib M, Ben-Mansour R, Badr HM, Ahmed SF, Ghoniem AF. Computational
combustion cases were investigated. The first six combustion cases
fluid dynamic simulation of oxyfuel combustion in gas-fired water tube
were based on the change of the out-of-service burners under full boilers. Comput &amp Fluids 2012;56(0):152–65.
load operation, while the rest cases were carried out at 20% lower [11] Eaton AM, Smoot LD, Hill SC, Eatough CN. Components, formulations,
and 20% higher than the standard operating conditions, as specified solutions, evaluation, and application of comprehensive combustion models.
Prog Energy Combust Sci 1999;25(4):387–436.
in Table 1. Temperature distributions, velocity distributions, spe- [12] Xu M, Azevedo JLT, Carvalho MG. Modeling of the combustion process and
cies concentrations, and char consumption are compared for all NOx emission in a utility boiler. Fuel 2000;79(13):1611–9.
examined combustion cases. The numerical results showed that [13] Belosevic S, Sijercic M, Oka S, Tucakovic D. Three-dimensional modeling of
utility boiler pulverized coal tangentially fired furnace. Int J Heat Mass Transf
the turned off burners in the opposite direction are better than 2006;49(19–20):3371–8.
those in the adjacent direction under full load operation, and this [14] Bosoaga A, Panoiu N, Mihaescu L, Backreedy RI, Ma L, Pourkashanian M, et al.
can help to provide an improved distribution of the flames in the The combustion of pulverised low grade lignite. Fuel 2006;85(10–11):1591–8.
[15] Díez LI, Cortés C, Pallarés J. Numerical investigation of NOx emissions from a
furnace. When the higher aerodynamic effects of the PC and feed tangentially-fired utility boiler under conventional and overfire air operation.
oxidizer gases were used, the combustion flames uniformly distrib- Fuel 2008;87(7):1259–69.
uted in the furnace zone. The predictions showed that there is a [16] Belosevic S, Beljanski V, Tomanovic I, Crnomarkovic N, Tucakovic D, Zivanovic
T. Numerical analysis of NOx control by combustion modifications in
clear reduction in the CO2 concentrations and char coal under pulverized coal utility boiler. Energy Fuels 2012;26(1):425–42.
the suggested operating combustion conditions in comparison [17] Tian ZF, Witt PJ, Schwarz MP, Yang W. Numerical modeling of victorian brown
with the reference case. A slight increase in the carbon monoxide coal combustion in a tangentially fired furnace. Energy Fuels
2010;24(9):4971–9.
concentrations was evident under the high and low load opera-
[18] Karampinis E, Nikolopoulos N, Nikolopoulos A, Grammelis P, Kakaras E.
tions. These predicted results were associated with the thermal Numerical investigation Greek lignite/cardoon co-firing in a tangentially fired
chemistry mechanism adopted in this study. Based on the above- furnace. Appl Energy 2012;97:514–24.
mentioned findings, the combustion characteristics and boiler per- [19] Dodds D, Naser J, Staples J, Black C, Marshall L, Nightingale V. Experimental
and numerical study of the pulverised-fuel distribution in the mill-duct
formance can be optimized for less emission. This is also important system of the Loy Yang B lignite fuelled power station. Powder Technol
to reduce the slagging and fouling phenomena on the heat transfer 2011;207(1–3):257–69.
surfaces of the furnace. [20] Al-Abbas AH, Naser J, Dodds D. CFD modeling of air-fired and oxy-fuel
combustion in a large-scale furnace at Loy Yang A brown coal power station.
Fuel 2012;102:646–65.
Acknowledgement [21] Anonymous: AVL Fire CFD Solver v 3.0 manual. A-8020 Gras, Austria,
www.avl.com; 2009.
[22] Al-Abbas AH, Naser J. Effect of chemical reaction mechanisms and NOx
The financial support from the Ministry of Higher Education modeling on air-fired and oxy-fuel combustion of lignite in a 100-kW furnace.
and Scientific Research in the Iraqi government (www.mohesr. Energy Fuels 2012;26(6):3329–48.
[23] Dukowicz JK. A particle-fluid numerical model for liquid sprays. J Comput Phys
gov.iq) is gratefully appreciated. 1980;35(2):229–53.
[24] Al-Abbas AH, Naser J. Numerical study of one air-fired and two oxy-fuel
combustion cases of propane in a 100 kW furnace. Energy Fuels
References
2012;26(2):952–67.
[25] Al-Abbas AH, Naser J, Dodds D. CFD modeling of air-fired and oxy-fuel
[1] Ahmed S, Hart J, Nikolov J, Solnordal C, Yang W, Naser J. The effect of jet combustion of lignite in a 100 kW furnace. Fuel 2011;90(5):1778–95.
velocity ratio on aerodynamics of a rectangular slot-burner in the presence of [26] Belosevic S, Sijercic M, Tucakovic D, Crnomarkovic N. A numerical study of a
cross-flow. Exp Thermal Fluid Sci 2007;32(2):362–74. utility boiler tangentially-fired furnace under different operating conditions.
[2] Hart JT, Naser JA, Witt PJ. Aerodynamics of an isolated slot-burner from a Fuel 2008;87:3331–8.
tangentially-fired boiler. Appl Math Model 2009;33(9):3756–67. [27] Agraniotis M, Nikolopoulos N, Nikolopoulos A, Grammelis P, Kakaras E.
[3] Ahmed S, Naser J. Numerical investigation to assess the possibility of utilizing a Numerical investigation of solid recovered fuels’ co-firing with brown coal in
new type of mechanically thermally dewatered (MTE) coal in existing large scale boilers – evaluation of different co-combustion modes. Fuel
tangentially-fired furnaces. Heat Mass Transf/Waerme Stoffuebertragung 2010;89(12):3693–709.
2011;47:457–69.