You are on page 1of 15

Powder Technology 171 (2007) 81 – 95

www.elsevier.com/locate/powtec

Numerical modelling of flow and transport processes


in a calciner for cement production☆
D.K. Fidaros, C.A. Baxevanou, C.D. Dritselis, N.S. Vlachos ⁎
Department of Mechanical and Industrial Engineering, University of Thessaly, Athens Avenue, 38334 Volos, Greece
Received 19 October 2005; received in revised form 1 September 2006; accepted 7 September 2006
Available online 29 November 2006

Abstract

Controlling the calcination process in industrial cement kilns is of particular importance because it affects fuel consumption, pollutant emission
and the final cement quality. Therefore, understanding the mechanisms of flow and transport phenomena in the calciner is important for efficient
cement production. The main physico-chemical processes taking place in the calciner are coal combustion and the strongly endothermic
calcination reaction of the raw materials. In this paper a numerical model and a parametric study are presented of the flow and transport processes
taking place in an industrial calciner. The numerical model is based on the solution of the Navier–Stokes equations for the gas flow, and on
Lagrangean dynamics for the discrete particles. All necessary mathematical models were developed and incorporated into a computational fluid
dynamics model with the influence of turbulence simulated by a two-equation (k–ε) model. Distributions of fluid velocities, temperatures and
concentrations of the reactants and products as well as the trajectories of particles and their interaction with the gas phase are calculated. The
results of the present parametric study allow estimations to be made and conclusions to be drawn that help in the optimization of a given calciner.
© 2006 Elsevier B.V. All rights reserved.

Keywords: CFD; Coal combustion; Calcination; Calciner modeling; Cement production

1. Introduction Dry heating of raw-mix in vertical suspension preheaters (see


Fig. 1) is mostly used, where calcination also takes place. The
The main processes of cement production include raw-mix innovation in the entire pyroprocess in modern cement plants is
preheating and calcination, clinker formation and cooling to the use of an additional calcining vessel, in which the raw-mix
achieve a crystalographic structure that meets the required ce- undergoes calcination to a level of 90 to 95%. In this way, the
ment specifications. After cooling, the clinker is fed into grind- calcined raw-mix enters the rotary kiln at a higher temperature,
ing or finish mills and is mixed with plaster and ameliorating thus reducing the energy demand and the thermal load on the
additives. The mills consume a very large amount of the total kiln. After being heated to the appropriate temperature, it enters
energy required for cement production. the calciner together with the fuel and the hot tertiary air, Fig. 2.
The raw-mix consists mainly of pulverized calcium carbon- The combustion heat released by the fuel causes calcination of
ate and silicon dioxide. During its heating/drying at tempera- the raw-mix according to the chemical reaction:
tures from 100 °C to 500 °C the moisture evaporates and at 850
1160 K
to 890 °C the endothermous calcination reaction begins, where CaCO3 Y CaO þ CO2 þ 178 kJ=mol ð1Þ
CaCO3 is converted into CaO and CO2. The activation energy
for the calcination is provided by the combustion heat of the The high fineness of the raw-mix and the good turbulent
fuel. mixing cause uniform and fast coal combustion and calcination
reactions. The products of the calciner are fed to the last cyclone
☆ that feeds the rotary kiln. The placement of calcination outside
Dedicated to the late Professor Shao-Lee Soo, for his pioneering work in
multiphase dynamics.
the cement kiln results in better quality of CaO and energy
⁎ Corresponding author. Tel.: +30 2421074094; fax: +30 2421074085. savings. For example, in the Olympus plant of AGET Hercules
E-mail address: vlachos@mie.uth.gr (N.S. Vlachos). in Greece calcination takes roughly 60% of the total heat
0032-5910/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.powtec.2006.09.011
82 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

while the higher rotational speeds can increase the quantity of


alkaline dust in the kiln, b) Reduction of NOx emissions is not
common in all cement production systems using calciners,
mainly due to geometric and operational differences, as well as
to different quality and quantity of raw-mix and fuels, and c)
The utilisation of fuels with low energy value, although eco-
nomically advantageous, requires particular attention in order to
avoid undesirable emissions of polluting and erroding gases.
From the above, it becomes apparent that control of cal-
cination is important because it affects fuel consumption,
pollutant emissions and the final cement quality. Therefore,
understanding the mechanisms of flow and transport phenom-
ena in the calciner may contribute to more efficient production
and better quality of cement.
Recently, calciners have been studied with different geom-
etries and operational conditions in 2D and 3D CFD simulations.
Huanpeng et al [2] studied the influence of various physical
parameters on the dynamics of gas–solid two-phase flow in a
precalciner using kinetic theory of granular flow to represent the
transport properties of the solid phase in a 2D model. Hu et al. [3]
used a 3D model for a dual combustor and precalciner using a
Eulerian frame for the gas phase and a Lagrangean one for the
solid phase in order to predict the burn-out and the decompo-
sition ratio during the simultaneous injection of two types of coal
Fig. 1. Schematic of cement production. and raw material into the device. Iliuta et al. [4] investigated the
influence of operating conditions on the level of calcination,
absorbed in the system, while 35% is spent for preheating and burn-out and NOx emissions of an in-line low NOx calciner, and
5% for clinkering [1]. This ratio of 60:40 is reversed in the case made a sensitivity analysis of their model with respect to aero-
where the calcination is taking place inside the rotary kiln. In dynamic and combustion/calcination parameters.
addition, the good mixing of fuel, air and raw-mix in the calciner In the present work a numerical model is described for the flow
results to faster calcination with good efficiency at relatively low and transport processes taking place in an industrial calciner. The
temperatures. model is based on the solution of the Navier–Stokes equations for
The advantages of using calcination devices are: a) The
addition of a burner in the calciner increases the capacity of the
rotary kiln in comparison to using simple preheaters, b) The
reduction of thermal load and the increased rotational speed of
the kiln (to achieve better mixing at increased capacity) extends
the lifetime of the firebricks and, thus, the operational life of the
kiln, c) The reduction of energy demand and the minimal calci-
nation in the kiln reduce considerably the exhaust gases and the
kiln heat losses to the environment because the exhaust gases
absorb most of the radiation, d) The combustion at medium–low
temperatures (b1400 °C) in the kiln reduces the production of
NOx, although combustion control and kiln burner design is still
significant, e) The lower temperature required in the calciner
allows the use of fuels with relatively low thermal capacity
(usually bituminous coal), f) The reduction of the thermal load of
the rotary kiln decreases the condensing of vapours (SO3, Na, K
and Cl) in the combustion area. However, the volatile cycle is
still a concern because now it will take place in the preheater/
precalciner tower itself as opposed to the kiln), and g) The
reduced calcification percentage in the rotary kiln, decreases its
thermal load and improves its functional stability, as the kiln
burners are now used only for clinkering.
Calciners have become essential devices in cement produc-
tion but have also disadvantages: a) The lower temperatures of
the exhaust gases may cause condensation of volatile alkalis, Fig. 2. Calciner device.
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 83

the gas flow and on Lagrangean dynamics for the discrete 2.4. Particle heat transfer
particles, using a commercial CFD code. All necessary flow, heat
and mass transfer and chemical reaction models are presented Particle heat transfer is due to convection, radiation and de-
with the influence of turbulence simulated by a two-equation volatilization, as follows:
(k–ε) model. Limited available measurements from the Olympus dm
cement plant of AGET Hercules are used to verify the model. h Ap Tl þ dtp hfg þ Ap ep rH4R
Tp ðt þ ΔtÞ ¼
h Ap þ Ap ep rTp3
2. Mathematical models 0 1
dm
h Ap Tl þ p hfg þ Ap ep rH4R
þ @Tp ðtÞ− dt A
2.1. Gaseous phase h Ap þ Ap ep rTp3
Ap ðhþep rTpe Þ
− t
The general form of the time-averaged transport equation for e mp Cp

momentum, heat and mass of the gases is: ð8Þ


  2.5. Devolatilization model
A A 1A 1 A W AU
ðqUÞ þ ðqU UÞ þ ðqrV UÞ þ q
At  Ax  r Ar
  r Ah  r Ah  The devolatilization model of Kobayashi [6] is used:
A AU 1A AU 1 A 1 AU  
¼ CU þ CU r þ CU þ SU E1
Ax Ax r Ar Ar r Ah r Ah R1 ¼ A1 exp ; ð9aÞ
ð2Þ RTp
 
where U, V, W are the time-averaged velocities in the axial, radial E2
R2 ¼ A2 exp ð9bÞ
and circumferential direction, respectively, ΓΦ the transport RTp
coefficient, and Φ any time-averaged transported fluid property.
where, R1 and R2 are competitive volatilization rates at different
temperature ranges. These yield an expression for devolatiliza-
2.2. Particle dynamics
tion:
Z t Z t 
The particle trajectories are calculated from their corre- mv ðtÞ
sponding motion equation: ¼ ða1 R1 þ a2 R2 Þexp ðR1 þ R2 Þdt dt ð10Þ
mpo −mash 0 0
dUp qp −q
¼ FD ðU−UpÞ þ gi þfi ð3Þ
dt qp
The Kobayashi model requires known kinetic parameters
where, the subscript “p” denotes particle. (A1, E1) and (A2, E2) and the contribution of the two reactions
For spherical particles, FD in the drag force term is: via the factors a1 and a2. More specifically A1 = 2.0e + 07 s− 1
3lCD Re and A2 = 1.0e + 07 s− 1 are the pre-exponential factors, and
FD ¼ ð4Þ E1 = 1.046e + 05 J/mol and E2 = 1.67e + 05 J/mol are the
4qp Dp2
activation energies. It is recommended that the value of a1
where the drag coefficient is calculated from: should be equal to the fraction of volatiles that is determined by
the proximate analysis, because this rate represents the volatile
a2 a3 evaporation at low temperatures. The value of a2 should be
CD ¼ a1 þ þ ð5Þ
Re Re2 equal to 1, as it expresses the contribution of the evaporation
rate of volatiles at very high temperatures.
and α1, α2 and α3 are constants proposed by Morsi and
Alexander [5]. 2.6. Surface/coal combustion models
The additional force term fi in Eq. (3) may be due to pressure
gradients, thermophoretic, Brownian or Saffman lift forces. After devolatilization is completed, there starts the surface
chemical reaction of the coal particle which may be modelled as
2.3. Particle size distribution follows:

The particle sizes follow a Rosin–Rammler distribution: 2.6.1. Diffusion model


n The reaction rate is determined by the diffusion of the gas
MD ¼ e−½ðD=D̄ o Þ  ð6Þ
oxidant into the particle surface:
where n is calculated from:
dmp mo Tp qg
lnðlnM D Þ ¼ −4pDp Dim ð11Þ
n¼ ð7Þ dt Sb ðTp þ Tl Þ
lnðD=D̄Þ
Each size interval is represented by an average diameter for In this model the particle diameter is assumed constant and,
which the trajectory calculations are performed. as its mass decreases, the active density decreases resulting in
84 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

a more porous particle. Eq. (11) proposed by Baum and The calculation of σp is repeated in the entire trajectory for n
Street [7] ignores the contribution of kinetics to the surface particles. Then, the source term that is introduced into the
reaction. energy equation is:
 
2.6.2. Kinetic/diffusion model rT 4
The reaction rate is determined by the diffusion of gas −jqr ¼ −4p a þ Ep þ ½a þ ap G ð19Þ
p
oxidant into the particle surface or by the reaction kinetics. The
model proposed by Baum and Street [7] and Field [8] is used, in 2.8. Chemical reaction models
which the diffusion rate is:
The present modelling of mixture fraction [11,12] with the
½ðTp þ Tl Þ=20:75
R1 ¼ C1 ð12Þ method of probability density function (mixture fraction/PDF)
Dp requires the solution of transport equations for one or two
conservative scalar properties. The effect of turbulence is also
and the kinetics rate: considered. The method of mixture fraction with PDF has been
  developed specifically for turbulent chemically reacting flow
E simulations. The chemical reaction is determined by turbulent
R2 ¼ C2 exp − ð13Þ
RTp mixing, which controls the limits of the kinetic rates. The PDF
method offers many advantages compared to the method of
The kinetics rate incorporates the effects of chemical reaction finite reaction rate. The method of mixture fraction allows the
in the internal surface of a coal particle and the epidermic explicit intermediate calculation of chemical compound form-
diffusion. The rates R1 and R2 are combined to give the ing and the interlacing of turbulence and chemistry. The method
combustion rate of the coal (char) particle. is economic, because it does not require the solution of a large
number of transport equations for each chemical species. More-
dmp R1 R2 over, it allows precise determination of auxiliary variables such
¼ −pD2p P0 ð14Þ
dt R1 þ R2 as density, and it does not use average values, in contrast to the
method of finite reaction rate.
The particle size is kept constant, until a significant reduction For a binary system such as fuel and oxidant, the mixture
in its mass leads to a new size estimation. fraction can be formulated in terms of elemental mass fractions:
Zk −ZkO
2.7. Particle radiation f ¼ ð20Þ
ZkF −ZkO
The radiation from the coal particles into the gas is The value of f is calculated from the solution of a time-
incorporated via the P-1 model [9–10]: averaged transport equation:
   
rT 4 A A A lt A f¯
jdðCjGÞ þ 4p a þ Ep −½a þ ap G ¼ 0 ð15Þ ðq ¯f Þ þ ðqui f¯ Þ ¼ þ Sm ð21Þ
p At Axi Axi rt Axi

where, Ep and αp are calculated from: The source term Sm is present only when particle mass
transport to the gaseous phase takes place.
X
N rTpn
4 Simultaneously with the solution of Eq. (21), a conservative
Ep ¼ lim epn Apn ð16aÞ f 2V , describing the
equation for the variance of mixture fraction, ¯
V Y0
n¼1
pV interaction between chemistry and turbulence, is solved:

X
N
Apn !
2V
ap ¼ lim epn ð16bÞ A ¯ A A lt A f¯
V Y0 V ðq f 2V Þ þ 2V
ðqui f¯ Þ¼
n¼1 At Axi Axi rt Axi
!2
The quantity Γ in Eq. (15) is: A¯f e 2V
þ Cg lt −Cd q f¯ ð22Þ
Axi k
1
C¼ ð17Þ
3½a þ ap þ rp  where, σt, Cg and Gd are constants equal to 0.7, 2.86 and 2.6,
respectively.
and σp is calculated from:
2.8.1. Coal reaction mechanisms
X
N Coal combustion — The most important physico-chemical
Apn
rp ¼ lim ð1−fpn Þð1−epn Þ ð18Þ change in the coal particle during heating is thermal frag-
V Y0 V
n¼1 mentation (pyrolysis) at high temperatures. During this stage an
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 85

important loss of weight occurs, because of dissolution of called metaplast. With the increase of temperature the metaplast
volatile matter, the quantity and composition of which depend is split, shaping the basic volatile products and semicoke,
on the ingredients of coal, its grain size and temperature. During causing the coal particles to swell. This is described by a factor
dissolution of volatiles, a number of parallel reactions occur, that depends on the composition of volatiles and the heating
with chemical combinations of reacting components or even rate. The increase of particle volume does not influence the
species such as, for example, CH4, CHOH, C2H6, H2, and S2. activity of pyrolysis, while the semicoke formed initially, is
After devolatization leading to production of water vapour, CO, decomposed as temperature increases.
CO2 etc, a series of progressive reactions of char and de- The rate of thermal decomposition increases with increasing
volatization gases take place as follows [1,13–23]: temperature up to a maximum value. Many researchers (for
Heterogeneous reactions example, [15,17,23]) have found that pyrolysis ends around 850
to 1000 °C, while its duration is limited to a few seconds
CðsÞ þ O2ðgÞ ⇒CO2ðgÞ ð23aÞ depending on the particle size. After the volatiles have been
released, the remaining solid (char) still retains a small per-
centage of volatiles (∼ 1.5%) like H2 and N2, requiring a tem-
2CðsÞ þ O2ðgÞ ⇒2COðgÞ ð23bÞ perature near 2000 °C to be removed completely.
Experiments show that the determination of volatiles in coal
is demanding and time-consuming. Many measurements of
CðsÞ þ 2H2ðgÞ ⇒CH4ðgÞ ð23cÞ volatiles based on the ASTM standard, present large differences
in the percentage of volatiles depending on the rate of tem-
perature increase and on the experimental method [10,15,16,
18,19]. The solid remains of the particles formed during thermal
CðsÞ þ CO2ðgÞ ⇒2COðgÞ ð23dÞ
decomposition are mainly fixed carbon, with high porosity and
large internal surface, and the inorganic part is ash. The tem-
perature varies between 1200 and 1800 °C causing ash melting.
CðsÞ þ H2 OðgÞ ⇒COðgÞ þ H2ðgÞ ð23eÞ The composition and the nature of ash as well as its properties
(melting point, viscosity, etc) depend to a large extent on the
Homogeneous reactions
pyrolysis conditions.
2COðgÞ þ O2ðgÞ ⇒2CO2ðgÞ ð24aÞ In cases where the gaseous phase consists mainly of air, the
pyrolysis and the combustion of char proceed simultaneously.
However, in general, char combustion follows pyrolysis, with
COðgÞ þ H2 OðgÞ ⇒CO2ðgÞ þ H2ðgÞ ð24bÞ only a very small time overlap. In ordinary coal particles, volatiles
tend to be emitted in concentrated but randomly distributed jets
from their surface. The larger jets reject volatiles during thermal
decomposition while smaller jets begin and end during this
COðgÞ þ 3H2ðgÞ ⇒CH4ðgÞ þ H2 OðgÞ ð24cÞ period. When the gaseous phase is hot enough and rich in oxygen,
the jets of volatiles ignite to form jet flames. In relatively large
particles, the emission and combustion of volatiles can keep the
CH4ðgÞ þ 2O2ðgÞ ⇒CO2ðgÞ þ 2H2 OðgÞ ð24dÞ char surface free of oxygen. When the surface of hot char is
accessed by oxygen, there begins a heterogeneous combustion
reaction with longer duration, lasting 15 to 20 times than the
HCðgÞ þ 1:5O2ðgÞ ⇒CO2ðgÞ þ H2 OðgÞ ð24eÞ thermal decomposition of volatiles, depending on its evolution
and combustion conditions [21–29].
The decomposition and polymerization reactions of the The heating rate of coal particles depends on their size and
superior and unsaturated hydrocarbons are also added: contact with the thermal source. For example, the heating rate of
coal powder by a surrounding flame is 1000 °C/s, but when the
fragmentation flame is from powder coal particles, the rate may increase to
Superior HCðgÞ Y Inferior HCðgÞ þ CðsÞ
10000 °C/s. The pyrolysis results in a number of products with
Unsaturated HCðgÞ YSaturated HCðgÞ large differences in molecular weight, from gaseous hydrogen
Polymerization up to heavy organic species (tar). The data provided by exper-
Unsaturated HCðgÞ þ H2ðgÞ Y Superior HCðgÞ iments concerning rapid pyrolysis is not sufficient to determine
the composition and distribution of intermediate products for
Pyrolysis — As temperature increases, the humidity and the various coals [30–32].
gases enclosed in the coal particles are released. The larger Thus, the mathematical models developed for devolatiliza-
percentage of the non-chemically combined water is evaporated tion are based on the initial coal particle composition. Many
at temperatures below 105 °C while the chemically combined at researchers, assume that the coal is considerably homogeneous,
temperatures exceeding 350 °C. At pyrolysis temperatures, so it is possible to be assumed as a heated mass and altered
certain types of coal melt, forming an intermediate product gradually from volatiles–char–ash to char–ash and finally to
86 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

ash. From tables of ultimate analyses of coal and pet coke [1,2], it for coal combustion and on the constitution and granulometry of
appears that the main components of volatiles are CO, CH, H2O, particles (average char diameter ≪100 μm), the selected model
and H2. Given that the atmosphere of the calciner is oxidant and for these particles was that of the kinetic/limited diffusion rate.
assuming that all these components react with oxygen, the main This is similar to that of shrinking-reactant particle core adopted in
reactions considered as taking place are: the general theory of surface heterogeneous chemical reaction.
The diffusion coefficient Dim of oxidant in the porous char used in
CO þ 1=2O2 →CO2 –283:2kJ=mol ð25aÞ the present model was 5.0e–05 m2/s.

H2 þ 1=2O2 →H2 O–242kJ=mol ð25bÞ 2.8.2. Calcination mechanisms


The calcination of limestone particles includes several stages,
CH4 þ 2O2 →CO2 þ 2H2 O–802:86kJ=mol ð25cÞ with each one imposing different chemical kinetics rates: a) Heat
transfer from the gases to the particle surface and from it to the
Char combustion — The mechanism of char combustion has reaction interface, b) thermal decomposition of CaCO3 in the
been investigated more than pyrolysis, without definitive reaction interface, c) mass flux of CO2 from the reaction in-
answers to questions concerning the quantitative origin of terface to the gases.
some constituents after the end of transformation. Qualitatively, For small limestone particles moving in high temperatures
however, it has been modelled satisfactorily by various mathe- gases, the internal and external heat and mass transfer rates are
matical models. These were developed in order to describe the high. Specifically, for particles with diameter between 1 and
solid coal combustion and have found important application in 90 μm and gas temperatures between 748 and 1273 K, Borgwardt
reactions of porous solids with gases. [39] has reported that the calcination is chemically controlled and
Two simple mathematical models describe the reaction of its rate is proportional to the surface area of the particle as
coal grain with oxygen: the simple film and the double film determined by the BET method (nitrogen absorption at 77 K).
model. In the first model the oxygen is diffused via a constant Because, the limestone microstructure is not completely crystalic
boundary layer in the surface of the char particle, where it reacts and has a diverse form of porosity, the surface determined by the
to form CO and CO2. The CO is then diffused in the well-mixed BET method, is the sum of the porous surfaces accessed by
environment. In the second model, char reacts with CO2 and not nitrogen. Under these conditions, the calcination happens on the
with oxygen, in order to produce CO that is burned in a thin total available surface, giving pseudo-volumetric characteristics
flame inside the boundary layer. The CO reacts with oxygen to the reaction.
inside the boundary layer, and thus the oxygen never approaches From the analysis of calcination data of high fineness
the char surface. Small particles (b 100 μm) are considered to limestone in isothermal reactors, it is concluded that, for a better
burn according to the first model and larger (up to N 2 mm) description of the reaction evolution, the model of shrinking core
according to the second. However, the two models constitute should be selected, with the size diameter raised to the power 0.6.
only the two extreme cases of char combustion and cannot, The value of the exponent (b 1) is explained by the fact that the
therefore, establish a general theory [29–36]. calcination proceeds radially to the particle core, without
The real mechanism of combustion is more complicated, inhomogeneities in the reaction interface. When the raw-mix
because of many factors involved such as particle size, local particles are small, the reaction interface of the calcination is not
temperature, local oxygen concentration and reaction controlling easy to determine. The internal thermal gradients and the partial
mechanism. Generally, the oxygen and CO are readily available pressure of CO2 are also difficult to estimate. For these reasons,
on the coal surface and can, therefore, react simultaneously with most calcination models of high fineness particles consider that
coal and also between each other. The situation becomes more the surface temperature is equal to the gas temperature,
complex when the char porosity is taken into consideration neglecting the internal thermal gradients [40].
(intrinsic model). The decomposition reaction of CaCO3 is strongly endother-
A more complex model proposed by Essenhigh [37] describes mic. Its thermodynamic state is defined by the reaction enthalpy
better the above processes. In this model the distribution of ΔH and the equilibrium pressure PCO2,eq,:
temperature and concentrations are extended to the center of the  
particle. The more usual diffusion controlled combustion of CO ΔH ΔS
PCO2;eq ¼ exp − þ ð26Þ
can be extremely fast, consuming all the local oxygen before it RT R
reaches the char surface and reacting only with CO2. In the
chemically controlled combustion of CO2 and O2, these have These values depend on temperature and are influenced by the
equal probability to react with the char surface. Moreover, ex- nature of limestone, its degree of cleanliness and mainly by its
perimental data by Field [8] and Borghi [38] showed that the structural mesh. The lower the degree of cleanliness of raw
reaction of char–CO2 is very slow in comparison with the reaction material, the lower is the reaction enthalpy. Also, the function of
of char–O2. Therefore, the latter can be considered as the main temperature–equilibrium pressure PCO2,eq = f(T) develops to
reaction on the char surface when the essential quantity of oxygen lower temperatures because of the chemical kinetics of the
is available. However, the presence of CO2 cannot be ignored and, recently formed CaO and the impurities in the reaction
thus, there always exists the probability of parallel reactions environment. The values of reaction enthalpies provided by the
[26,27,29–31]. Based on a comparative analysis of existing data open literature [40–46] for the present endothermic reactions
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 87

Fig. 3. Variation of calcination chemical kinetics with temperature.

vary. Particularly, in the high interest range for calcination (800 to


1000 °C), the reaction enthalpy is not linearly dependent on
temperature. Thus, for practical calculations ΔÇ900 = 1660 kJ/kg
CaCO3 = 396 kcal/kg CaCO3 can be assumed a constant value for
this specific temperature range [1,41].
The decomposition of limestone takes place in a reaction
zone, where the core of unreacted CaCO3 and the newly formed
CaO meet. This front moves from the perimeter to the center Fig. 4. Calciner side view.
with a certain speed, while heat is transferred simultaneously to
the core and CO2 is emitted to the outside. This reaction The reaction rate at the interface is expressed as follows,
proceeds in the following stages: a) Heat is transferred from the Borgwardt [39]:
surroundings to the particle surface, b) heat is conducted
Rate ¼ −ks ACaCO3 ð28aÞ
through the reacted layer to the reaction zone, c) chemical
reaction occurs in the reaction zone, CO2 emission, nuclei where,
creation and reforming of CaO, and d) CO2 is diffused through  
Ea
the CaO layer to the particle surface and the surroundings. ks ¼ Aexp − ð28bÞ
RT
The final reaction speed is a function of the rates of the
above stages. Because these rates are of the same order of The activation energy Ea of the decomposition reaction is in
magnitude, a balance is achieved in the decomposition front, the range 165–205 kJ/mol.
under the prevailing temperature and partial pressure of CO2, The calcination of small limestone particles dispersed in the
so that the rates of the above stages become equal. If large gaseous phase, can proceed at temperatures up to 1600 °C. The
limestone particles exist, diffusion of mass and conduction of effect of CO2 partial pressure is incorporated in the decompo-
heat will dominate, especially when the surrounding temper- sition rate by modifying it as proposed by Darroundi and Searcy
ature is high and the partial pressure low. In the case of low [48]:
temperatures and high partial pressures of CO2, the material
transformation occurs by the diffusion of CaO. For a fine ksV ¼ ks for Pb10−2 Pe ð29Þ
granulometry of ground limestone or raw-mix in ordinary
calcination conditions, the chemical kinetics play a decisive ksV ¼ ks ðPe −PÞ=Pe for 10−2 Pe bPbPe ð30Þ
role [1].
Thus, the proposed model, calculates the rates of particle The effect of temperature on calcination chemical kinetics is
calcination and heat transfer by considering: a) the heat transfer shown in Fig. 3. During calcination, the thermal conductivity of
by convection from the gases to the particle and by conduction
to the particle interior, b) surface decomposition of CaCO3, and Table 1
c) mass transport of CO2 from the reaction interface via the Mass flow rates at the inlets of the calciner
porous particle to the gaseous environment. Kind of Case 1— coal Case 2 — pet coke
The calcination is a heterogeneous reaction and occurs at the mass flow
Quantity [kg/s] Percentage Quantity [kg/s] Percentage
lime surface when the local pressure exceeds the criterion of mCaCO3 52.47 54.3% 52.47 54.6%
Baker [47]: mCoal 3.78 3.9% 3.17 3.3%
  mTertiary Air 39.36 40.7% 39.36 41.0%
19; 680 mAir Coal 0.97 1.0% 0.97 1.0%
Pe ¼ 1; 826  10 exp −
7
ð27Þ
T Total 96.59 100.0% 95.98 100.0%
88 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

Table 2 The physico-chemical processes take place in the main volume


Ultimate analysis for the solid feeds of the calciner, consisting of a 6.6 m diameter cylinder with
Components Case 1— coal Case 2 — pet coke 20 m height. The upper conical part has 1.1 m height and leads
As used (%) Dry basis (%) As used (%) Dry basis (%) to a cylindrical part with 4.3 m diameter and 5 m height. The
Humidity 1.45 0.0 1.45 0.0 total calciner volume is 850 m3. The coal entries are at 2.4 m
Volatiles 27.70 28.11 13.21 13.40 height from the start of the cone and at 2.68 m from the
Ash 12.05 12.23 0.14 0.14 calciner axis.
Fix carbon 55.80 56.62 81.70 82.90 The computational domain consists of a hybrid mesh of
Total 97.00 96.96 96.50 96.44
67.104 cells. Because of symmetry, the calculations were carried
out for one half of the calciner using the FLUENT code. Two
lime depends on the state of the solid material and differs con- fuels (coal and pet coke) were considered and the total rate of
siderably for non-calcinated, partially calcinated or fully mass (raw-mix, coal and air) fed into the three kinds of inlets was
calcinated particles. This is mainly due to the different structure, aproximately 100 kg/s. As shown in Table 1, the larger per-
but also to the change of the specific surface area. In the reaction centage of mass rate is that of CaCO3, followed by tertiary air,
region, the thermal conductivity of the particle is a linear coal and finally the coal feeding air.
function of specific surface area and temperature. Thus, the The Rossin–Rammler distribution of the raw-mix size had
thermal conductivity of CaCO3 was 1.646W/(m.K) and of CaO an average value of d = 16.6 μm and a spread parameter of
0.860 W/(m.K). The mass fraction of CO2 is determined from a n = 0.822, while the coal had d = 34.5 μm and n = 1.248. The
diffusion equation assuming a spherical particle. analysis of the raw meal and coal particles is given in Table 2.
The tertiary air entered with a velocity 24 m/s, coal with 11.5 m/s
3. Computational details and the raw-mix with 1.5 m/s. The coal was fed pneumatically
while the raw-mix entered by gravity.
3.1. Calciner geometry All the geometric data and the initial and boundary conditions
were supplied by Olympus plant of AGET Hercules in Volos,
The modeled calciner, Fig. 4, consists of a cylindrical and a Greece.
conical section having three kinds of inlets at the bottom part
and an outlet at the top, from where the products such as 4. Results and discussion
calcined raw-mix, CO2, and other gases exit. Raw-mix is fed
into the calciner via two 0.6 m diameter pipes inclined at 60° 4.1. Case 1 (Good quality coal)
to the horizontal. The tertiary air enters axially from the bottom
via a concentric 2.6 m diameter duct and the coal is fed at the Fig. 5 shows the calculated velocity distribution of the gaseous
lower conical part via two 0.2 m pipes at 30° to the horizontal. phase for Case 1 (good quality coal) in two vertical diametral

Fig. 5. Velocity distribution in vertical symmetry plane (left) and at 90° (right) for Case 1.
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 89

Fig. 6. Temperature field in the vertical symmetry plane (left) and at 90° (right) for Case 1.

planes normal to each other. The gases undergo an abrupt little above the threshold for calcination, so that calcification
deceleration at the beginning of the lower conical part due mainly takes place almost in the whole device. There are no spots of
to the entry of the coal and secondly of the raw meal. In the main high temperature but rather regions of low temperature because
cylindrical part, the velocity remains at 7 to 8 m/s, with regions of of intense calcination, resulting to high heat absorption. The
higher velocity in front of the two raw-mix inlets and in the upper high temperatures in regions where higher velocities prevail are
conical part. At the exit, a region with higher velocities is mainly due to the high concentration of burning coal and to the
observed, a fact due to the relative absence of particles. absence of CaCO3 particles.
In Fig. 6 higher temperatures are observed in the opposite Fig. 7 shows concentration distributions of CO2, O2 and H2O
side of the raw-mix entries. This is due to the trapping of small in various horizontal cross-sections. Higher concentrations are
coal particles while the concentration of CaCO3 particles is low. observed a little after the raw-mix inlet. It should be noted that
The main body of the calciner is maintained at temperatures high CO2 concentrations result to high heat absorption, thus

Fig. 7. Concentration distributions of CO2 (left), H2O (middle) and O2 (right) for Case 1.
90 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

limiting calcination. The CO2 concentration decreases progres-


sively towards the exit, mainly because the available CaCO3 is
also decreasing. The amount of CO2 produced by coal com-
bustion is much smaller than that due to the calcination (ratio
1:6). Smaller O2 concentrations are observed in regions of high
temperature, mainly due to pyrolysis and the start of combustion
of the coal remains (char). In the part of the cylinder where
intense calcination takes place, the O2 concentration remains at
low levels because of the high raw-mix and CO2 concentrations,
while in regions where coal particles have been trapped, lower
CO2 concentrations are observed (mainly from coal combus-
tion). Higher H2O concentrations are recorded after the fuel inlet
in the regions of pyrolysis and coal combustion. These
concentrations decrease along the height of the device and
away from the combustion region.
The trajectories of coal particles are presented in Fig. 8 and
those of CaCO3 in Fig. 9. The average passage length covered by
Fig. 9. Trajectories of CaCO3 particles for Case 1 (calcination is marked in red).
all particles is about 54 m, while the longest exceeds 75 m. The
right parts of Figs. 8 and 9, show the parts of the trajectories where
the particles are activated thermochemically. Thus, the blue colour deceleration, again for the same reasons as in Case 1. In the
in the start of the trajectory corresponds to coal warming up, the main cylindrical section of the calciner, a region with higher
green and yellow to evaporation and combustion of volatiles, velocities opposite to the two raw-mix inlets and in the upper
respectively, and the red to the combustion of fixed carbon. conical part is observed. The gas velocity reaches 8.5 m/s,
Finally, the blue colour in the end of the trajectory shows the encouraged by the relatively small particle load. In contrast, the
cooling of the ash by the gases. The average passage length of coal velocities decrease to 5–6 m/s in the remaining part. This is
particles is 32 m, while the longest exceeds 45 m. Their average mainly due to the higher particle concentration, which
residence time is 5 s, mainly because the air moves faster influences the exiting speed of the gases.
sweeping the smaller coal particles. The average particle In Fig. 11 higher temperatures relative to Case 1 are observed.
residence time is 10 s while the longest 15 s. Calcination is This is attributed to the better quality and utilization of the fuel.
noticeable in the semi-cylinder defined by the raw-mix feeding Higher temperatures are observed in the opposite side of the raw-
pipes and the coal inlets, which agrees with the temperature field mix inlets, mainly due to the trapping of small coal particles
and the CO2 concentration. The predicted calcination for Case 1 while the concentration of CaCO3 particles is exceptionally
reaches 96.5%, and is realised in all the active calciner height. small. The main body of the calciner is kept at temperatures well
above the effective calcination temperature, so that large CaCO3
4.2. Case 2 (Pet coke) quantities are calcinated in short times. The high temperatures
seen in regions where higher speeds prevail are attributed to the
Fig. 10 shows the gas velocity field for Case 2 (pet coke absence of CaCO3 particles that would consume the heat re-
fuel). In the start of the conical part, the gases undergo a strong leased. Also at the exit, higher temperatures are observed, be-
cause calcification there is exceptionally limited.
Fig. 12 shows that higher concentrations of CO2, O2 and
H2O are present at a small distance after the raw-mix inlet.
Again it should be noted that high CO2 concentrations result to
high heat absorption, as far as it does not limit the calcination.
CO2 concentration decreases gradually along the 16.5m calciner
height until the exit, where the available quantity of CaCO3 is
also decreased. Smaller O2 concentrations are observed in the
regions where the temperatures are high (pyrolysis and com-
bustion of solid coal remains). Also, in the semi-cylindrical part,
where intense calcination takes place, the concentration of O2
remains at low levels because of high raw-mix and CO2
concentrations, while in regions where coal particles have been
trapped, still lower concentrations are observed. Higher con-
centrations of H2O are seen little after the coal entry and in the
pyrolysis and coal combustion regions. These concentrations
are decreasing along the calciner height.
From the predicted trajectories of pet coke particles (not
Fig. 8. Trajectories of coal particles for Case 1 (combustion is marked in red). shown), most calculated residence times did not exceed 10 s
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 91

Fig. 10. Velocity distribution in vertical symmetry plane (left) and at 90° (right) for Case 2.

with the longest reaching 16.2 s. The average trajectory length water vapour levels. This results from the volatiles and
for all particles (raw-mix and pet coke) was roughly 52 m, while hydrogen compounds in the char. The fast combustion of pet
the longest exceeded 73 m. The average residence time of pet coke forces the ash to abandon the calciner faster than other
coke particles was smaller (4.5 s) than in Case 1. The average particles.
trajectory length of these particles was 30 m while the longest From the predicted trajectories of CaCO3 particles (not
42 m. It should be noted that the combustion of most pet coke shown), their average residence time reached 10 s and the longest
particles is completed inside the device. The higher tempera- 15.6 s, indicating that the speed of these particles is higher than
tures observed are due to the intense and fast char combustion, that of the pet coke. The longer residence time corresponds to the
and are accompanied by low concentrations of O2 and high particles that collide in the upper conical part (before the exit) and

Fig. 11. Temperature field in the vertical symmetry plane (left) and at 30° (right) for Case 2.
92 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

Fig. 12. Concentration distributions of CO2 (left), H2O (middle) and O2 (right) for Case 2.

are trapped by the rising particles from the central part of the
calciner. Calcination is realised fast and in a region close to the
calciner axis. The calcination rate for the particular fuel (Case 2)
reaches 98.7%, without taking advantage of the total active height
by the majority of CaCO3 particles. The reason that a small raw-
mix quantity is not being calcinated is the large diameter of
CaCO3 particles and the rapid acceleration observed near the
device exit.
Fig. 13 shows details of the fuel particle trajectories. It is
evident that the pet coke particles (Case 2) react faster than those
of coal (Case 1), as soon as they enter the calciner.
Finally, the evolution of CaCO3 calcination for Cases 1 and 2 is
depicted in Fig. 14. The differences relate to the energy prevailing
levels observed, the aerodynamics (mass density, particle load)
and the CO2 concentrations that suppress calcination. The
evolution of calcination is represented by the CaCO3 decompo-
sition along the calciner height, starting from where calcification

Fig. 13. Details of fuel particle trajectories for Case 1 (upper) and Case 2 (lower).
(The pet coke particles (Case 2) react faster than the coal particles (Case 1)). Fig. 14. Evolution of calcination with height for Cases 1 and 2.
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 93

is initiated until the device exit. The active calcination height for chemical activation and NOx formation. Small recirculations
Case 1 is approximately 27.5 m. An amount of 50% of CaCO3 is are also formed and act as a bumper wall for the particles,
calcinated in the first 6 m, while the calcification rate is stabilized resulting to local high temperatures and erosion of the walls.
to lower values for the next 16 m and is reduced in the last part, Finally, the model was capable to predict the range of velo-
where gas acceleration is also observed. cities and gas temperatures and the asymmetric distributions in
The active calcination height for Case 2 is also 27.5 m with the calciner exit, as observed in the Olympus cement plant of
differences observed mainly in the evolution of the calcification. AGET Hercules.
An amount of 70% of CaCO3 is calcinated in the first 7.5 m,
while the calcination rate gradually decreases with height. In the Nomenclature
last 5 m, the calcination rate decreases further, due mainly to the A pre-exponential factor
increasing gas speed, despite the fact that the temperature is well Apn area of projection of n-particle
above the calcination threshold. Fig. 14 shows also the Ap area of projection of a particle
suppressive role of increased CO2 concentration, which is of ACaCO3 CaCO3 concentration
great importance for the performance of the calciner. In the a1, a2 devolatilization coefficients or evaporation factors
region defined by the lower conical part and the raw-mix inlets, CD drag coefficient
the temperature should be maintained at a specific range in order C1,2 constants
to avoid locally high CaCO3 production. This is related to the Dp particle diameter
CO2 partial pressure that constitutes the basic controlling D0 size constant
parameter of calcification. Dim diffusion coefficient of the oxidant
Although there are no detailed measurements to compare d diameter
with the present numerical results, it should be noted that the D̄ mean diameter
model is capable to predict the range of gas temperatures (850– E, Ea activation energy
900 °C) and the asymmetric distribution of the particles (coal Ep equivalent particle brightness
and raw-mix) in the exit of the calciner, as observed in the epn brightness of the nth particle
Olympus cement plant. In Case 1 (coal fuel), the velocities at FD coefficient for drag force term
the exit of the calciner are in the same range as those measured fi additional force
(17–21 m/s) with higher values observed in regions of higher fpn scattering factor of nth particle
temperatures. There were no data available to compare the f mixture fraction
numerical results for Case 2 (pet coke fuel). G incident radiation
gi gravitational acceleration
5. Conclusions hfg evaporation latent heat
h thermal convection coefficient
A numerical model was presented for the prediction of the k thermal conductivity
velocity, temperature and concentration fields of gases and of kS surface reaction rate
particle trajectories in an industrial low NOx calciner for cement MD mass fraction of particles with diameter larger than D
production. Models were also included for radiation, chemical mv(t) sum of volatiles evaporated up to time t
reactions and turbulence effects. From the results of the param- mpo initial particle mass
etric study the following main conclusions can be drawn: mash mass of ash in the particle
The small recirculation region observed near the raw-mix and mo local fraction of gas oxidant
tertiary air inlets, increases the active length of the device and the mp particle mass
particle residence time. The rapid calcination near the raw-mix inlet N total number of particles in a volume V
produces high local CO2 concentrations, which limits calcination. n size distribution parameter
Higher temperatures are observed near the coal inlet where P pressure
combustion of volatiles occurs. The high temperature regions Pe local pressure
observed along the calciner are due to coal particles with orbits P0 partial pressure of oxidant in the gas environment of
that are not intermingled with the raw-mix, suggesting that more particle
attention should be paid to its granulometry than to that of coal. PCO2,eq equilibrium partial CO2 pressure
The high unevenness of the Rossin–Rammler particle dis- qr radiation flux
tribution used is mainly responsible for the small percentage of R universal gas constant
the non-calcinated CaCO3. Despite the energy surplus observed Re Reynolds number
(more in Case 2 than Case 1) calcination could not be completed R1 surface reaction rate
because of the high CO2 concentration released immediately R2 surface devolatilization rate
after the entry. Sb stoichiometric coefficient of the reaction
The upper conical part of the calciner causes an acceleration Sm source term
of the gases and particles, which reduces calcination. The T temperature
resulting high temperatures can cause problems of thermal Tp particle temperature
stresses, and can lead to erosive by-products from SiO2 Tpn temperature of n-particle
94 D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95

t time [10] M.Q. Brewster, T. Kunitomo, The optical constants of coal, char and
U gas velocity limestone, ASME Journal of Heat Transfer 106 (1984) 678–683.
[11] Y.R. Shivatahnu, G.M. Faeth, Generalized state relationships for scalar
Up particle velocity properties in non-premixed hydrocarbon/air flames, Combustion and
U,V,W time-averaged axial, radial and circumferential Flame 82 (1990) 211–230.
velocities [12] W.P. Jones, J.H. Whitelaw, Calculations methods for reacting turbulent
zk mass fraction for the chemical element k flows: a review, Combustion and Flame 48 (1982) 1–26.
[13] R.A. Srinivasan, S. Sriramulu, S. Kulasekaran, P.K. Agarwal, Mathemat-
zKF mass fraction for fuel stream
ical modeling of fluidized bed combustion — 2: combustion of gases, Fuel
zKO mass fraction for the oxidizer stream 77 (1998) 1033–1049.
ΔS change of grammolecular entropy [14] R.L. Backreedy, R. Habib, J.M. Jones, M. Pourkashanian, A. Williams, An
ΔH reaction enthalpy extended coal combustion model, Fuel 78 (1999) 1745–1754.
[15] F. Winter, M.E. Prah, H. Hofbauer, Temperatures in a fuel particle burning
Greek letters in a fluidized bed: the effect of drying, devolatilitazion and char
combustion, Combustion and Flame 108 (1997) 302–314.
α absorption gas factor [16] A.S Hull, P.K. Agarwal, Estimation of kinetic rate parameters for coal
αp equivalent absorption factor combustion from measurements of the ignition temperature, Fuel 77 (1998)
ΓΦ transport coefficient 1051–1058.
εp particle brightness [17] L.D. Smoot, International research center's activities in coal combustion,
ΘR particle temperature produced by the intensity of Progress in Energy and Combustion Science 24 (1998) 409–501.
[18] L. Huilin, Z. Guangbo, C. Rushan, C. Yongjin, D. Gidaspow, A coal
thermal radiation combustion model for fluidized bed boilers, Fuel 79 (2000) 165–172.
μ gas molecular viscosity [19] L. Huilin, Z. Gungbo, B. Rushan, Y. Lidan, Q. Yukun, Modeling of coal
μt turbulent viscosity combustion in a 25-MW FBC power plant, Energy 24 (1999) 199–208.
ρp particle mass density [20] H.H. Liakos, K.N. Theologos, A.G. Boudouvis, N.C. Markatos, Pulverized
ρg gas density coal char combustion: the effect of particle size on burner performance,
Applied Thermal Engineering 18 (1998) 981–989.
σ Boltzmann's constant [21] W. Sujanti, D. Zhang, A laboratory study of spontaneous combustion of
σp particle scattering equivalent factor coal: the influence of inorganic matter and reactor size, Fuel 78 (1999)
Φ time-averaged transported fluid property 549–556.
[22] C.N. Eastwick, S.J. Pickering, A. Aroussi, Comparisons of two com-
Acknowledgements mercial computational fluid dynamics codes in modeling pulverized coal
combustion for a 2,5 MW burner, Applied Mathematical Modeling 23
(1999) 437–446.
This work was partially supported by the General Secretariat [23] L. Armesto, J.L. Merino, Characterization of some coal combustion solid
for Research and Technology of Greece and the cement residues, Fuel 78 (1999) 613–618.
company AGET Hercules (EPET-II/96SYN121). The authors [24] M.S. Khan, S.M. Zubair, M.O. Budair, A.K. Sheikh, A. Quddus, Fouling
resistance model for prediction of CaCO3 scaling in AISI 316 tubes,
wish to thank Dr. I. Marinos and Mr. T. Pissias of AGET for the
International Journal of Heat and Mass Transfer 32 (1996) 73–79.
data and for useful discussions. [25] J.R. Fan, X.H. Liang, Q.S. Xu, X.Y. Zhang, K.F. Cen, Numerical simulation
of the flow and combustion processes in a three-dimensional, W-shaped
boiler furnace, Energy 22 (1997) 847–857.
References
[26] J.K. Fink, Pyrolysis and combustion of polymer wastes in combination with
metallurgical processes and the cement industry, Analytical and Applied
[1] E. Kolyfetis, C.G. Vagenas, Mathematical modeling of separate line Pyrolysis 51 (1999) 239–252.
precalciner, ZKG International 2 (1988) 559–563. [27] P. Basu, Combustion of coal in circulating fluidized-bed boilers: a review,
[2] L. Huanpeng, L. Wentie, Z. Jianxiang, J. Ding, Z. Xiujian, L. Huilin, Chemical Engineering Science 54 (1999) 5547–5557.
Numerical study of gas–solid flow in a precalciner using kinetic theory of [28] F.C. Lockwood, T. Mahmud, M.A. Yehia, Simulation of pulverized coal test
granular flow, Chemical Engineering Journal 102 (2004) 151–160. furnace performance, Fuel 77 (1998) 1329–1337.
[3] Z. Hu, J. Lu, L. Huang, S. Wang, Numerical simulation study on gas–solid [29] R.G. Bautista-Margulis, R.G. Siddall, L.Y. Manzanares-Papayanopoulos,
two-phase flow in pre-calciner, Communications in Nonlinear Science & Combustion modeling of coal volatiles in the freeboard of a calorimetric
Numerical Simulation 11 (3) (2006) 440–451. bed combustor, Fuel 75 (1996) 1737–1742.
[4] I. Iliuta, K. Dam-Johansen, A. Jensen, L.S. Jensen, Modelling of in-line- [30] C.A. Gurgel, J. Saastamoinen, J.A. Carvalho, M. Aho, Overlapping of the
low-NOx calciners — a parametric study, Chemical Engineering Science devolatilization and char combustion stages in the burning of the coal par-
57 (2002) 789–903. ticles, Combustion and Flame 116 (1999) 567–579.
[5] S.A. Morsi, A.J. Alexander, An investigation of particle trajectories in two- [31] W.B. Fu, B.L. Zhang, S.M. Zheng, A relationship between the kinetic
phase flow systems, Journal of Fluid Mechanics 55 (1972) 193–208. parameters of char combustion and the coal's properties, Combustion and
[6] H. Kobayashi, J.B. Howard, A.F. Sarofim, Coal devolatilization at high Flame 109 (1997) 587–598.
temperatures, Proc. 16th Intl Symposium on Combustion, The Combustion [32] T. Gentzis, F. Goodarzi, C.N. Koykoyzas, A.E. Foscolos, Petrology,
Institute, 1976, pp. 411–425. mineralogy and geochemistry of lignites from Crete, Greece, Coal Geology
[7] M.M. Baum, P.J. Street, Predicting the combustion behavior of coal particles, 30 (1996) 131–150.
Combustion Science and Technology 3 (1971) 231–243. [33] J. Zhang, S. Nieh, Comprehensive modeling of pulverized coal combustion
[8] M.A. Field, Rate of combustion of size-graded fractions of char from a in a vortex combustor, Fuel 76 (1997) 123–131.
low rank coal between 1200K–2000K, Combustion and Flame 13 (1969) [34] L. Reh, Challenges of circulating fluid-bed reactors in energy and raw
237–252. materials industries, Chemical Engineering Science 54 (1999) 5359–5368.
[9] S.S. Sazhin, E.M. Sazhina, O. Faltsi-Saravelou, P. Wild, The P-1 model for [35] H. Wang, J.N. Harb, Modeling of ash deposition in large-scale combustion
thermal radiation transfer advantages and limitations, Fuel 75 (3) (1996) facilities burning pulverized coal, Progress in Energy and Combustion
289–294. Science 23 (1997) 267–282.
D.K. Fidaros et al. / Powder Technology 171 (2007) 81–95 95

[36] I.E. Sarris, O. Giannopoulos, D. Fidaros, N.S. Vlachos, Modelling coal [43] F. Acke, I. Panas, Activation energy of calcination by means of a temperature
combustion in a rotary cement kiln, 3rd National Congress on Computational programmed reaction technique, Thermochimica Acta 306 (1997) 73–76.
Mechanics, Volos 24–26 (June 1999) 665–672. [44] J. Khinast, G.F. Krammer, C. Brunner, G. Staudinger, Decomposition of
[37] R.H. Essenhigh, in: M.A. Elliot (Ed.), Chemistry of coal utilization, Second limestone: the influence of CO2 particle size on the reaction rate, Chemical
Supplementary Volume, Wiley, New York, 1981, pp. 1153–1178. Engineering Science 51 (1996) 623–634.
[38] G. Borghi, A.F. Sarofim, J.M. Beer, 70th AIChE Annual Meeting, New York, [45] N. Hu, A.W. Scaroni, Calcination of pulverized limestone particles under
vol. 34c, Nov. 1977, pp. 14–17. furnace injection conditions, Fuel 75 (1996) 177–186.
[39] R.H. Borgwardt, Calcination kinetics and surface area of dispersed limestone [46] S.V. Krishnan, S.V. Sotirchos, Effective diffusivitty changes during
particles, AIChE Journal 31 (1) (1985) 103–111. calcination, carbonation, recalcination and sulfation of limestones,
[40] F. Acke, I. Panay, Activation energy of calcination by means of a temperature Chemical Engineering Science 49 (1991) 1195–1208.
programmed reaction technique, Thermochimica Acta 300 (1997) 73–76. [47] E.H. Baker, Calcium oxide-carbon dioxide system in the pressure range
[41] V.L. Boris, Mechanism of thermal decomposition of alkaline-earth 1–300 atmospheres, Journal of Chemical Society (London) (1962) 464–470.
carbonates, Thermochimica Acta 303 (1997) 161–170. [48] T. Darroudi, A.W. Searcy, Effect of CO2 pressure on the rate of decomposition
[42] A.K. Galwey, M.E. Brown, Arrhenius parameters and compensation of calcite, Journal of Physical Chemistry 85 (26) (1981) 3971–3974.
behavior in solid-state decompositions, Thermochimica Acta 300 (1997)
107–115.

You might also like