This article was downloaded by: [Moskow State Univ Bibliote]

On: 08 February 2014, At: 06:23

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Combustion Science and Technology

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/gcst20

Mathematical Simulation of a Circulating Fluidised Bed

Combustor
a a a
P C. SARAIVA , J. L. T. AZEVEDO & M. G. CARVALHO
a
Mechanical Engineering Department , Institute Superior Técnico , Lisbon, Portugal
Published online: 24 Oct 2007.

To cite this article: P C. SARAIVA , J. L. T. AZEVEDO & M. G. CARVALHO (1993) Mathematical Simulation of a Circulating
Fluidised Bed Combustor, Combustion Science and Technology, 93:1, 223-243, DOI: 10.1080/00102209308935291

To link to this article: http://dx.doi.org/10.1080/00102209308935291

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the
Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Combust. Sci. and 'tech; 1993, Vol. 93, pp. 223·243 @Gordon and Breach Science Publishers S.A.
Photocopying permitted by license only Printed in the United States of America

Mathematical Simulation of a Circulating Fluidised Bed

Combustor

P c. SARAIVA, J. L. T AZEVEDO and M. G. CARVALHO Instituto Superior Tecnco,

Mechanical Engineering Department, Lisbon, Portugal

Abstract-In this paper a mathematical model to describe a CirculatingAtmospheric Fluidised Bed Combustor
(CAFBC) is presented. For the fast section of the bed, momentum and energy balance equations are used
to predict temperature and velocity profiles for gas and particles. The model performs mass balances for the
chemical gas species (02, H20, CO2. CO and S02) with consideration on the last being given for retention
by limestone particles. A bubbling bed model is considered to simulate the bottom of the CAFBC. The
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

model is applied to typical conditions of a boiler and the results show the expected trends.

NOMENCLATURE
A -Area (rn")
C -Concentration (moles)
Cp -Specific heat Capacity (J/KgK)
d -Particle diameter (m)
Dg -Diffusion coefficient (m 2/s)
H -Bottom region height (m); Heating Value (J/Kg)
h -Convection heat transfer (W/m 2K)
kc, kd -Kinetic rate; Diffusion rate (s -I)
m -Particle mass (Kg)
P -Pressure (Pa)
T -Temperature (K)
u -Velocity (m/s)
m -Rate of combustion (Kg/m/s)
W -Mass flux (Kg/m/s)

Greek

8 -Density (Kg/m")
e -Voidage; Emissivity
JL -Viscosity (Kg/ms)
P -Bulk density (fuel and limestone) (Kg/m3 )
p; -Density of a chemical specie (K~i/m3.;,)
o -Stefan Boltzman constant (W/m K4 )

Subscripts
o -Initial conditions
f -fuel particles
g -gas
I -limestone particles
mf -minimum fluidization
p -Particle
rad -radiation
s -inert particles
v -volatiles
w -wall

223
 224 P. C. SARAIVA, J. L T. AZEVEDO AND M. G. CARVALHO

INTRODUCfION

Circulating f1uidised bed boilers have a number of advantages over conventional bubbling
f1uidised bed boilers and became an important topic of research. The empirical design of
boiler furnaces practiced for many years promoted studies to understand the processes
occurring in such equipment. Nowadays, due to restrictive pollution regulation as well as
the necessity of high performance and the development of powerful computers, research
programs were created.
Investigations on high-velocity fluidization have been conducted by many investigators
(see e.g., Yerushalmi and Avidam, (1985), Kwauk et al., (1987), Rhodes and Geldart,
(1987». These models allowed the identification of the circulating f1uidised bed structure,
with a more dense zone at the bottom and a fast bed above, with higher particle
concentrations near the wall region. Kunni and Levenspiel (1990), developed a model
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

for the hydrodynamics which considers three different phases, (dilute particles entrained,
ascending and descending clusters), with mass transfer between them. The key parameter
in their model is the decay constant for the fall off of bulk density of solids with height
in the freeboard. Berruti and Kalogerakis (1989), modelled the fast bed hydrodynamics
using a core-annulus theory which is applied to small diameter risers (d < 0.3 m) used
in catalytic processes.
Wei(J et al. (1987) developed a mathematical model which describes coal combustion
in a CAFBC, based on mass and energy balances for a compartment and considers
eight chemical species. However, no details are given about the solids distribution in
the reactor and no special treatment is given to the bottom region. At the bottom of a
circulating f1uidised bed there is a dense phase which has particle concentrations typical
of a bubbling f1uidised bed (see e.g. Kwauk et al., (1987), Kunni and Levenspiel, (1990».
A compilation of all relevant work on the modelling of circulating f1uidised beds and
conventional bubbling beds is presented by Manno and Rietema (1990),
The present paper proposes a model for a Circulating Atmospheric Fluidised Bed
which includes hydrodynamics of the fast bed as well as for a bubbling bed. Combustion,
heat transfer and S02 retention are modelled throughout the reactor (for details see
Saraiva (1992». After this introduction the mathematical model for the fast and bubbling
bed are described, followed by the model of S02 retention by the limestone particles.
The model is applied to operating conditions of a typical boiler and the results are
discussed. In the last section the conclusions of the present work and future work for
the model are outlined.

MATHEMATICAL MODEL

Fast Bed

The model used for the fast bed region of the combustor was adopted from Chung and
Carlson (198\) and extended to a two-zone core-annulus model. In the core, particle
velocities are obtained by integration of the equations of motion. When only particle
drag and gravitational forces are considered the particle momentum equation can be
written as:
du; I
up - = -(u g - up) - g (1)
dx Tp

where Tp is the particle's relaxation time expressed as:

(2)
 MATHEMATICAL SIMULATION OF A CIRCULATING FLUIOISED BED COMBUSTOR 225

The drag coefficient is calculated using the well known correlation

24
CD = - ( 1 + 0, 15Re p0,687) 1'-4.67 (3)
Rep
The gas phase velocity is evaluated by integration of the momentum equation:

d~ 1 0 1 ~
ug - = -(u g- ur)- + -(ug - us)- - (up - ug) Sburn (4)
dx Tr Pg Ts Pg
where momentum exchange between the gas and particles phase is considered and
including the source of momentum exchange due to coal combustion, Sburn.
For combustion calculations the energy equations for both phases are used to predict
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

the temperature profiles. The evolution of coal particles is assumed to be in three steps.
In the first one, coal particles are dried (in this step temperature is assumed constant and
equal to 1000 C) and heated until the devolatisation temperature. Then devolatisation
starts and this process is kinetically controlled. When devolatisation is complete, char
combustion is considered to be controlled by both kinetics and oxygen diffusion to the
particle's surface. The energy equation for a particle is based on mass and energy balances
and can be written as:

dTr . ...4 4 dmH,o

Cpr mr (j( = A r mv,e (H v.e -Qv)+A r h(Tg-Tr)+ArO"cr (lg -Tr )+HH,O~ (5)

where Qv represents the fraction of the particle heat of combustion which is transferred
to the particle and m, is the mass flow rate released:

. Or R, E;
m, ="3 [1 - ( Rr ) 3] RrAvexp ( )
- RTr (6)

for volatile combustion. For char consumption the mass flow rate by combustion, me, is
expressed by:
• < 2
me = 1r k, vr drPo , (7)
where
1 1 1
-=-+- (~
k, kc k<t
kc and kd represent the kinetic and diffusion rates. The main assumptions done are:
i) the particles are spherical ii) a one-step reaction for combustion iii) the particle
density remains constant iv) particles are sufficiently separated from each other that the
single-particle combustion analysis is valid for each and v) the temperature of the particle
is uniform.
The energy equation for inert particles is similar to equation (5) but without considering
the drying, volatilization and combustion steps (mv,e = 0).

dT s 4 4
Cps mS(j( = As h(Tg - T s ) +A,O"cs(Tg -Ts ) (9)

To predict gas phase temperature, an energy balance is performed on a small compartment

of height 6x, leading to:
 226 P. C. SARAIVA, J. L. T. AZEVEDO AND M. G. CARVALHO

d
dx (PgUg CPgTg + PC U, (CpC Tc + Hv,c - O«) + Ps Us Cps Ts) =
4h
O(Tw - T g ) + Srad + Qrad (10)

where D is the combustor diameter. The radiative heat fluxes to the combustor walls
as well as radiative sources for the gas phase, are evaluated by the Discrete Transfer
Method. The overall emissivity is calculated considering the gas as a mixture of grey gases
and particles obeying to the Geometrical Optics Theory. The convection coefficient, h,
is predicted according to the model of Mahalingam and Kolar (1991) for circulating
fluidised beds.
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

The mass concentrations of fuel and chemical species are expressed as a function of
the mass combustion rate:

d 3m v c pc
-(pc u.) = - - - '- - a(pc uc) (lla)
dx ocRc
d
-(Ps us) = -a(ps us) (lIb)
dx

(1Ic)

d Meo, . d[CO)
dx (peo, ug) = C2 Me mv.c + Mco, -d-t- (lid)

d Mco . d[CO)
-(peo ug) = C3 - - mv.c - Meo-- (lie)
dx Me dt

(I If)

(llg)

where the constants C 1 to C s are evaluated from stoichiometry, The last term of equation
(1Ia) and (lib) represents the radial diffusion of particles from the core to the annulus.
The value of the constant "a" was taken from Rhodes and Geldart (1987). The CO
oxidation rate used in equation (lid) was taken from Howard et al. (1973) correlation
and the last term of equation (1Ig) represents the S02 retention by limestone.
The gas density is related with temperature to a given pressure by the equation of
state:

(12)

Bubbling Bed

The model adopted for the bubbling bed is a "Level 11" model classified as a "learning
model" by Van Swaaij (1985). The gas flow is divided into two phases--emulsion and
bubble. Mass transfer between the bubble and particle phases is evaluated by a mass
 MATHEMATICAL SIMULATION OF A CIRCULATING FLUIDISED BED COMBUSTOR 227

coefficient kbe , proposed by Sit and Grace (1981), considering diffusive and throughflow
effects:

11'.]
[2U mr + 12 (Dg~:Ub)
Kt,.=-=--------=------=-
Db
(13)

where Db is the bubble diameter predicted by a correlation established by Mori and

Wen (1975). For a heterogeneous reaction, with negligible particle concentration in the
bubble phase, it is assumed that no reaction occurs in the lean phase. Assuming uniform
concentration in the particulate phase Cp, the gas concentration in the bubble phase C,
can be expressed along the height as:
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

Cb = c, + (Co - Cp)exp (_ ~ecbX) (14)

/3U g
Considering a heterogenous first order reaction with rate k, in the particulate phase, an
overall balance is:

Ug Co = /3 Ug CbH + (1 - /3) Ug Cp + kR (1 - cb) CpH (15)

The overall rate of carbon consumption may then be expressed as

kR =L 7r d~ k r; F; Cchar (16)

Carvalho et al. (1991) established the reaction rate as:

1 1 dr
(17)
kr = kc + ¢ Sh' D:.,
where Sh' is the Sherwood number:

Sh' = [4+ 0.576 (U o, d r)0078 +1 028(Uo,dr) +00141(d r) (Uo,d r

C Dmc Dmc d, Dmc
)3]
1/2 (18)

and the kinetic rate k, is given by La Nauze (1985);

k, = 0.513 Tr exp (-9160/Tr) (19)

Equations (14) and (15) are solved together to calculate CbH (x = H) and Cpo The
parameter zi of equations (14) and (15), as well as bubble fraction, Cb, must be calculated
from bed hydrodynamics. A modified two phase theory is used (e.g. Azevedo (1991))
considering the through flow in bubbles (n=3) and the gas velocity in the emulsion is
assumed to be larger than the minimum fluidization velocity by a I<e factor correlated
for (cb < .3) as:

(eb > 004) (20)

The particulate phase voidage is assumed to remain equal to the incipient fluidization
voidage and the bubble velocity is obtained according to Davidson and Harrison (1963):

Ub = u, - Umr + 0.711 Jg Db (21)

 228 P. C. SARAIVA, J. L T. AZEVEDO AND M. G. CARVALHO

Considering the above mentioned modified two-phase theory the bubble void fraction is
given by:

U, - K.,U rnf
Ch = :-:---:-:-----:---=---==---::--:- (22)
U b + Urnf(n + 1 - K.,)
The fraction of flow within the bubbles, (3, is predicted by:

Urnf K.,
(3 = 1- - - ( 1 - ch)
Uo
(23)

The objective of the dense region model is to predict the char concentration in the bed
for each class of particles considered and also the mass flow rate due to splashing. Wen
and Chen (1982) correlated the entrainment flux of solids at the bed surface as:
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

9 ITD 2 p~.5 g0.5

Eo = 3 x 10- Db(U o - Urnf)- 25 (24)
4 /lg'

so the mass flux for each size of particles is a function of carbon concentration and mass
fraction of each diameter:

(25)
To predict the mass fraction of each class (Fi) a mass balance for each size is performed:

Falling flux
rfrom fast bed
1- r Ejecting flux from
bubbling bed
1r + Flux from i + 1 size
by combustion
1-
Flux from i-I size] _ [Fraction of combustion] =0
[ by combustion particles burnt
which can be written as:

W f , lI; - E, + W i +1 - Wi - W hum , = 0 (26)

The first term of the above equation is the particle flux of the falling particles related
to equations (1) and (1Ia). The mass flux due to combustion is directly proportional to
the reaction rate (see equations (15) and (16)):

(27)
It is assumed that, from the initial number N of particles of ilb size, Ni-I will change to
i_1 1h size and N, will remain with its original dimensions. The corresponding mass flux
from a given size to the smaller one is:

_ H Obed C char (d~ - di) (dfi_I ) 3

Wi_I -
r,
(3
dfi - d
3) - dfi . F, (28)
fi_ 1
For the smallest class Wi = 0 and for the largest W i + 1 = O. In equation (28) dr is the
particle diameter after the combustion process during a mean particle residence time
(t, ):
IP
_ d3 6 krt r
df f - - (29)
- ( IT dr Cchar )
 MATHEMATICAL SIMULATION OF A CIRCULATING FLUIDISED BED COMBUST,R 229

Equation (26) can be re-written in the form:

Ai F; + B; Fi+1 =C i (30)
forming a set of algebraic linear-equations. The char concentration (C ch. , ) is calculated
from an overall mass balance of the bubbling bed. The bed height is a result of the
mass balances of the limestone particles in all the combustor. Since bulk density in the
fast bed is calculated, the total mass of solids in the upper region of the combustor is
a result of the integration over the total height. Knowing the total amount of solids in
the combustor, the dense region height is then calculated.

S02 Retention
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

During coal combustion, the sulfur compounds are oxidized and the resultant sulfur
dioxide is reduced by calcium oxide particles, (produced by the limestone's calcination),
forming calcium sulfate according to the reaction:
I
S02 + Ca 0 + 2 02 --+ Ca S04

The reaction rate of a limestone particle can be expressed as (Borgwardt, (1970)):

11" 3
kl = 6 d, kvl CSO, (31)

In equation (31), kV1 represents the overall volumetric reaction rate constant and Cso,
is the S02 concentration in the combustion gases. The overall volumetric reaction rate
is calculated by:

(32)
where Sg is the specific surface area correlated with calcination temperature (Borgwardt
et al (1971)) given by:

Sg = -38.4 T + 5.6 X 104 T ~ I253K (33a)

Sg = 35.9 T + -3.67 X 104 T < I253K (33b)

and Al is the limestone's reactivity which is a function of the fractional conversion of
CaO, temperature and particle size.

NUMERICAL PROCEDURE
The set of non-linear differential equations governing momentum and energy in the fast
bed are solved using the Runge-Kutta method. Equations (lla) to (llg) are manipulated
by the introduction of equation (IIa) into equations (llc) to (IIg). By direct integration,
the gaseous species are calculated in each compartment. For S02 retention, the reaction
rate is evaluated at all the compartments and mass balances are performed.
In the bubbling bed, the set of algebraic linear equations written in the form of equation
(30) is solved directly assuming the char concentration in the bed. Thus an overall mass
balance is performed to predict the new char concentration and an iterative procedure
is established.
The initial conditions for the fast bed are taken from the feed conditions and from
the solids splashing from the bubbling bed. For the bubbling bed calculation, the flux of
 230 P. C. SARAIVA, J. L. T. AZEVEDO AND M. G. CARVALHO

TABLE I
Operating Conditions
Conditions Mass Flow
Air A ; 1.2
Fuel (Kgh- I) 480
Additives (Kgh -I) 15.3

TABLE II
Coal Composition
Coal
Lower Heating Value (MJ/Kg) 9.85
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

H 20 (%) (raw) 53.60

Ash (%) (raw) 3.2

Volatil Matter (%) 49.40

C(%) (waf) 66.30
H(%) (waf) 4.76
S(%) [waf] 0.46
N(%) (waf) 0.97
0(%) (waf) 27.20

incoming particles is the sum of the falling flux of solids and the downflow of particles
near the walls of the fast bed.
The coal particles are considered in a discrete number of sizes which are tracked along
the fast bed. height. Each particle size is partially burned out in each passage in the
combustor and its diameter is thus reduced. The recirculated particles are considered
in the size classes closer to the diameter at the end of the previous passage. The size
classes include diameters down to the size which are completely burned in one passage.
The procedure considered allows the consideration of only two kinds of coal particles
along the combustor, the ones in the first passage which are still releasing moisture and
volatiles and the recirculated ones with char combustion.

DISCUSSION OF RESULTS
The model described in the previous section was applied to a typical circulating fluidised
bed boiler. The overall dimensions are 9,4 m high, and a cross-sectional area varying
from 0,16 m2 (at the distributor) to 0,42 m2 (at 3,1 m height and above). The combustion
air is supplied through the distributor (primary air) and at the secondary air inlets at
0,7 m, 2,7 m, 5,5 m and 7,1 m.
Most of the energy released by the combustion of air and fuel is radiated to the
walls covered by water tubes. It is assumed that the boiler's walls have tubes at uniform
temperature. In the present work the bubbling bed temperature was fixed at 1123 K. The
operating conditions and coal composition used are shown in tables 1 and 2, respectively.
The results of the bubbling bed section are not analysed in detail in the present work.
This module was introduced in the calculation procedure to allow for the calculation of
the gas and reacting particles' initial conditions for the fast bed.
The outlet gas concentrations from the bed are shown as the initial conditions for the
fast bed section after dilution with the secondary air (24.8% of the total). The downflow
 MATHEMATICAL SIMULATION OF A CIRCULATING FLUIDISED BED COMBUSTOR 231

0.30

1--

0.24
Z
0
......
E-<
;:) 0.18
I!l
......
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

0::
E-<
(f)
......
0 0.12
~
N
......
(f)

0.06

0.00
600. 900. 1200. 1500. 1800. 2100.
Dp [urn]

FIGURE I _Bed; O-Feed

solids near the walls (and calculated from equations (1Ia) and (Llbj), together with the
falling particles, is the total feed rate to the bubbling bed.
The char in the bed was found to be about 2.8% of the total weight. The particle size
distribution of the feed coal and the resultant in the bed are represented in Figure I.
The projected solids have a particle size distribution similar to the one existing in the
bed and the initial particle velocity was taken as twice the bubble velocity.
The predicted temperature profiles along the fast bed of the combustor are presented
in Figure 2. The inlet temperature of feed, splashed and recirculated coal particles is
about 373 K, 1100 K and 700 K, respectively. The feed coal particles are dried and heated
in the initial zone of the fast bed leading to a slight decrease in the gas temperature
until coal particles ignite. The volatiles released from the coal particles falling in the bed
were assumed to be burned instantaneously at the bed surface. For the splashed and
recirculated particles, only char combustion is considered since drying and devolatisation
are complete in the first passage in the combustor and during the residence time in the
bed. Due to these particles, the gas temperature remains almost constant along the whole
combustor, which is typical for CFBe. The figure also shows that the drying time is not
negligible and takes about 20-25% of the particle's residence time in the first passage
on the combustor.
Figure 3 represents particles' and gas velocities. The gas velocity evolution is a direct
consequence of the gas temperature and gas/particle interactions. The drift velocity is
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

ti
N

Tg
Op=0.600

1800.00 ___ Op=1.100

_ _ _ Op=O.740 '"o
:;;a
.......
1350.00 ~ _ _ _ _ Op=0.900
/'
Pt:l
,-<,--==~=--~_:-:.:::-~=~~
,. , ... -....--- - - ..-- =--~, ~
~ <;.....,<"'" ;:: - _ _ _ _ Op=0.600 ~
t> ,
"
1./ r
/ ./ ~

E-< ( ________ Op=1.000 r

900.00 ~ / // ,.;
-ex: ,(
~
/ I >
Pt:l f;j
0.. <
~ I-
I / g
Pt:l 450.00 _-, __ J
E-< ~
1=
I I I
p

2.4 4.8 7.1 9.5 12

~
~
:r
X lml o

FIGURE 2 Temperature profiles. The dashed lines are representing fresh (T ra = 100 K), recirculated (Tra
= 770 K) and projected (T fo = 1123 K) particles for different dimensions.
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

Vg ;;::
~
10.00 · I :I:
Dp=D.6DD m
;;::
___ Op=1.lOD ~
-~------- Q
r-
........ _ _ _ Op=D.740 Vl
en 7.50 I-
<, ~
C
-:~--:::~--=~~:~-~-~~~~~~:-~:=-:=-:_::_~_?_~~~~ _ _ _ _ Dp=D.9DD
~/ ---;.
.....8 1/ ---<: - ~
o
II ,j _ _ _ _ Op=D.600 z
>-
E-< 5.00 I- 1,'; o-n
..... /
________ Dp=1.DDD ;,.
U
o o
....:l r'
~
8
> 2.50 I- ~
z
o
r
0.00' , , , , ,I , , ! , I! , , , , I, , , ! t !
E
8
V;
g)
0.0 2.4 4.8 7.1
'"c:Jm
X lml 8;;::
'"e
FIGURE 3 Velocity profiles. The dashed lines are representing fresh (de = 600 I'm and 1100 I'm and 900 a
I'm). ;>l

!;j
w
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

Ii!
__ _ [)P=O.7 40
100.00 .-
___ [)P=1.100
.- .- .-
.- .- ___ [)p=O.600
...- ...- .-
75.00 I- ./
-- Dp=0.900 '"o
...... /'
~ /' - --- :>
'"
...... Dp=0.740
./7':'--
E-<
.,;-- ___ Dp=1.100
~
p
::::> ,...
o 50.00 ~ -- --
/ / -- r
-- --/ '" -- ,.,
/ /
"'.::..- - -- - :>
--I-~ --
~
I- gJ
i 25.00 /
1//
'"
- - - ----
- o
1/ 1z
~/
.A
-- -- •
-- • I I
"
~
0.00 o
2.4
---./7------ -- 4.8 7.1 9.5 ~
0.0
"r-~
X lml :t
o

FIGURE 4 Coal Panicles Burn-Out: fresh (d f = 740 I'm and 1100 I'm), projected (dr = 600 I'm and 900
I'm) and recirculated (dr = 740 I'm and 1100 I'm)
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

:::
~
:t
Gwall m
:::
200.00 __ Gconv ~
§
~
r-
cz _ _ _ Grad en
i:C
150.00 I-
~ ~
~ oz
::0::::
~ o-rt
CI:l :I>
~ 100.00 I- r:
>< ~C
:::> I-
~
~ ~-- ~
"""'--.~ Z
E-< I- --~ o
50.00 - ~ ~
<I:: - ~~-
~ - ..... ~
~
:J::: - -- 8
enm
I- ---~---I----I---- o
0.00 m
'"
'='
0.0 2.4 4.8 7.1 9.5 8:::
X lml 'e"
§
;ll

FIGURE 5 Heat Flux to the Walls; Radiative Heat Flux to (about 70% of the total); Convective Heat Flux
(about 30% of the total)
tl
~
 236 P. C. SARAIVA. J. L. T. AZEVEDO AND M. G. CARVALHO

higher for the feed particles leading to a high initial acceleration allowing the dynamic
equilibrium to be reached in the first two meters of the combustor. For the smallest
diameters, higher velocities are reached and the limestone particles have considerably
smaller velocities due to their larger density.
Figure 4 represents profiles of predicted burnout, for the coal entrained from the
feed, splashed and recirculated, function of particle diameter. It is clear that the fresh
particles are not totally consumed, especially the larger ones. The residence time in the
fast bed is very small in one passage. However, typical particles in a CAFBC make tens,
or even hundreds, of circuits before leaving the system, with each circuit requiring 2
to 20 seconds in the riser (fast bed) and minutes in the return system providing a high
residence time. The inclusion of a recirculation model shows that recirculated particles
have higher burnout levels depending on their dimensions (50% for 1100 J.tm and 85%
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

for 740 I,m). These recirculated particles represent all the passages of coal particles after
the initial one. In these passages only the char combustion is considered which allows
a common treatment for particles with different residence times. The total degree of
burnout is dictated mainly by the cyclone efficiency which is considered in this test with
11 = I. However, the highest burnout levels are reached by the projected particles from
the bottom region, since they have dried and released volatiles before.
Heat flux profiles to the walls are shown in Figure 5. It can be seen that 70% of the
total heat flux to the walls is by radiation transfer while the other 30% is transferred by
convection. Since the operating conditions are typical of low loading, the bulk density is
not very high. In this case the gas/particle mixture is not dense, allowing the radiation
to play an important role in the heat transfer process.
Figure 6 shows concentration profiles of COz, CO, HzO and Oz. In the present model
it is assumed that the reaction rate is directly proportional to the reaction rate of coal
combustion. The high concentrations of O 2 (10% for 20% of excess air) and low levels
of CO 2 emissions show that there is less combustion in the upper part of the combustor
than in the lower one. This result is a consequence of the value "a" in equations (11a)
and ( II b) since it has been evaluated for different operating conditions. The content of
water vapour is high in the flue gases due to the high moisture content in the lignite
considered.
The influence of limestone particle size and Ca/S molar ratio on the S02 retention
was studied. Figure 7a) shows S02 concentration profiles with height. The S02 profile
without limestone retention is represented in curve 1 in the combustor (fast bed) for
the case of molar ratio Ca/S equal to two and particle sizes of 150 I,m and 500 J.tm,
respectively. Figure 7a) shows that more S02 is retained by the larger particles in the
first passage. To analyse these results, Figure 7b), is helpful. Figure 7b) represents the
limestone's reactivity which is a function of the fraction of CaO not reacted to form
CaS04. The smallest particles have higher reactivities than the largest particles which
means a larger capability to absorb S02 in the combustion gas, but their residence time in
the fast bed is lower. However, considering recirculation, the residence time of particles
becomes similar for all particle sizes. Thereby, the retention of S02 is higher for lower
limestone diameters due to the higher reactivity for a given conversion fraction.
Figure 8a) shows the S02 emission profiles for various Ca/S ratios and a fixed particle
size (d, = 250 I,m). It is clear that S02 emissions decrease with increasing Ca/S ratio.
Figure 8b) shows that the reactivity of limestone does not remain similar along the
combustor (for the three ratios used), showing the importance of mass flow rate in the
fast bed. Decreasing the calcium to sulphur ratio, the conversion of CaO to CaS04
within the particles is increased
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

:s:
~
::t
02
:s:'"
0.40 __ CO2 ~
§
r-
/'
-- ___ COxlO en
....., 1 §2
/
C
0 - I I ___ H2O
:> 0.30 --------,
-
<, /
\
E
<5
z
/
0 -----1
I o."
-
~
:>
/ \ ------ >
0.20 - o
Eo-< I
Z 8
~
U ~
z
z c
0 0.10
u ,
~ -- -----'------ ~
/ L - - - ---- in
"gJ
0.00 /' - , ~ - - -1- - - - -
-I - - - - - -I - - - - - -
m
'"e
0.0 2.4 4.8 7.1 9.5
8:s:
X lml '"c
~
;C

FIGURE 6 OZ. COZ. CO and HzO profiles along the combustor with air staging.

tJ
....
 238 P. C. SARAIVA, J. L. T. AZEVEDO AND M. G. CARVALHO
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

'"
II

a
'0
~
~
~

.2
s
:~ [
UJ8
o~
'"
0

0 0 0 o. 00
0 0 0 . 0 0
0
0
-.:!'
0
0
C"':l
0
0
N

[A A] Wdd
-
0
0
0
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

1 :::
1.00 :!i
:t
1 rn
:::
___ 2 :!i
§
___ 2
0.75 f-
._---~~;:::~--~:-------::~. ...en
~---------------- 3:
C
>< ______ 3
Eo-<
~
;;> ______ J sz
Eo-< 0.50 f- o."
----
u :l>
<
~
(")

~
f-
~
0.25 ~
Z
o
."
r-
I I I S
o
0.00 v;
0.0 2.4 4.8 7.1 9.5 ~
'mo"
X lml 8:::
'c"
FIGURE 7b) Reactivity of Limestone Particles for the Case of Ca/S = 2; I - dl = 500 I'm; 2 - d t = 250
em; 3 - d, = 150 I'm.
~:<l

s
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

400.00 ___ 2
___ 3
____ .4
300.00
~
'"o
t> ~
........
:>
~
s:
200.00 ~
~
::s
0-. r
,.,
10-. >
N
m
100.00 /' '""""'- - - - -- - - ---. -- -- ---- <

f,// - - - '- - - _ _ _ _ 8
--:::----- .... _--- ------------- ~
0.00' t c. , , , , I , , ! , , , , , ! , r - ,- -,
~
p
0.0 2.4 4.8 7.1 9.5
~
~
X lml
~
o

FIGURE 8a) S02 Emission for the Case of d, ; 250 I'm: 1 - OIJS ; 0; 2 ; CaJS ; 0.5; 3 - CaJS ; 1; 4 - CaJS ; 2.
 Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

:::
1.00 ~
:t
m
:::
-- ------ 2 ~
------------------- 2 ~r-
0.75 -------~-~-~-~---------~-~-~-~- on
>- - - -- ______ 3 i:
C
E-<
...... --
:> ______ 3 ~
(3
...... --- Z
E-< 0.50
u o."
< ;,>
~ ()
p:::;
~
C
0.25
~
51
o
."
r-
C
0.00 aVi
0.0 2.4 4.8 7.1 9.5 m
til
"m
X lrnl "8
ztil
c
FIGURE 8b) Reactivity of Limestone Particles for the Case of d t = 250 I'm: 1 - CaiS = 0.5; 2 - CatS = I; 3 - CaiS = 2. §
'"
'i:
 242 P. C. SARAIVA, J. L. T. AZEVEDO AND M. G. CARVALHO

SUMMARY
A numerical model to simulate a two zone regime with combustion in the fast bed
of the CAFBC was implemented. This model was coupled, for the first time, with
a model for the bottom region of the combustor derived from bubbling bed theory.
The model allows for the calculation of gas and particle velocities, gas concentration,
temperature and heat fluxes, along the reactor. A model for S02 retention was also
included.
The model includes some coefficients which are difficult to calculate and therefore
empirical values have to be considered. Validation of this model, which is still
required, will be performed when experimental results of the furnace considered
become available.
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014

ACKNOWLEDGEMENTS

This work is currently supported by the Commission of the European Communities under the contract 0030-C
(MB) entitled "Minimization of the Formation of Air Pollutants in the CAFBC by using European Fuels
and Additives" of the JOULE subprogram of Solid Fuels. The author P. C. Saraiva wants to express his
gratitude to JNICT for the scholarship given for Msc.

REFERENCES

Azevedo. J. L. T., (1991). On the influence of bubble fraction and velocity on bed permeability, 1ST internal
report.
Bcrruti, F. and Kalogerakis, N., (1989). Modelling the internal flow structure of circulating fluidised bed,
The Can. 1. of Chern. Eng.• 67, pp. 1010-1014.
Borgwardt. R. H., (1970). Kinetics of the reaction of S02 with calcined limestone, Env. Sci. Techn., 4, pp.
49-53.
Borgwardt, R. H., Drehmel, D. C, Kittleman, T. A., Mayfield, O. R. and Owen, J. S., (1971). Selected
studies on alkaline additives for sulphur dioxide control, EPA-RTP.
Breault, R. W.• Mathur, V. K.• (1989). High velocity fluidised bed hidrodynamic modelling, Ind. Eng. Chern.
Res, 28, pp. 684-688.
Carvalho. J. R. F. G., Pinto, M. F. R. and Pinho. C M. C T, (199\). Mass transfer around carbon panicles
burning in fluidised beds, Trans. IChem. E.. 69, pp. 63-70.
Chung, P. M. and Carlson, L.. (1981). Vertical combustion for particulate refuse, Numerical Heat Transfer, 4,
pp. 101-122.
Davidson, J. F., and Harrison, D., (1963). Fluidised Panicles. Cambridge, Cambridge University Press, pp.
21.
Howard, J. 8., Williams. G. C and Fine, D. H.• (1973). Kinetics of Carbon Oxidation in Post Flame Gases,
14th Symp. (Int.) on Combustion. The Combustion Institute, pp. 975-986.
Kunni, D. and Levenspiel, 0., (1990). Entrainment of solids from fluidised beds, Powder Technology, 61, pp.
193-206.
Kwauk, M., Ningde. w.. Youchu. L.. Bingyu, C, and Zhiyuan, S.• (1987). Fast fluidization at ICM, Circulating
Fluidised Bed Technology, pp. 33-45.
La Nauze, R. D., (1985). Mass transfer considerations in fluidised bed combustion with particular reference
to the influence of system pressure, Chern. Eng. Res. Des., 63, pp. 219-229.
Mahalirngam, M. and Kolar, A. K., (1991). Emulsion Layer Model for Wall Heal Transfer in a Circulating
Fluidised Bed, AIChE 1., 37, pp. /139-1150.
Manno. V. P., Rieterna, S. M., (1990). An annotated bibliography of fluidised bed combustion modelling
information. Powder Technology. pp. 23-34.
Mort, S. and Wen, C Y., (1975). Estimation of bubble diameter in gaseous fluidised beds, AIChE 1.. 21, pp.
109·115.
Rhodes, M. J., Geldart, D., (1987). A model for the circulating fluidised bed, Powder Technology, 53, pp.
155-162.
Sara iva, P. C. (1992). Modelacao de Combustao. Transferencia de Calor e Emissao de Poluentes em Caldeiras
de Leito Fluidizado Circulante, Msc Thesis (in Portuguese).
Sit, S. P. and Grace. J. R., (1981). Effect of bubble interaction on interphase mass transfer in fluidised beds,
Chemical Eng. Science, 36, pp. 327-335.
Van Swaaij, J. R.• (1985). Chemical Reactors. Fluidization. Academic Press, pp. 595-629.
Wei/1, V., Scholer. J. and Feu. F. N., (1988). Mathematical modelling of coal combustion in a circulating
fluidised bed reacting, Circulating Fluidised Bed Technology II, pp. 289-298.
 MATHEMATICAL SIMULATION OF A CIRCULATING FLUIDISED BED COMBUSTOR 243

Wen, C. Y. and Chen, L. H., (1982). Fluidised bed freeboard phenomena: Entrainment and elutriation,
AIChE J., 28, pp. 117-1287.
Yates, J. G., (1983). Fundamentals of ftuidised bed chemical processes, Butterworths.
Yerushalmi, J. and Avidam, A., (1985). High Velocity Fluidization, Fluidization, Academic Press, pp. 225-291.
Downloaded by [Moskow State Univ Bibliote] at 06:23 08 February 2014