Modeling of clinker cooler

8/9/2019 Modeling of clinker cooler

1/18

Chemical Engineering Science 62 (2007) 2590 2607www.elsevier.com/locate/ces

Rotary Cement Kiln Simulator (RoCKS): Integrated modeling ofpre-heater, calciner, kiln and clinker cooler

Kaustubh S. Mujumdara ,b, K.V. Ganesha, Sarita B. Kulkarni a, Vivek V. Ranadea,

aIndustrial Flow Modeling Group, National Chemical Laboratory, Pune 411 008, IndiabDepartment of Chemical Engineering, Indian Institute of TechnologyBombay, Powai, Mumbai 400 076, India

Received 10 January 2007; accepted 26 January 2007

Available online 14 February 2007

Abstract

This paper presents an integrated reaction engineering based mathematical model for clinker formation in cement industry. Separate models

for pre-heater, calciner, rotary kiln and cooler were initially developed and coupled together to build an integrated simulator. Appropriate models

for simulating gassolid contact and heat transfer in pre-heaters were developed. Calciner was modeled by considering simultaneous combustion

of coal particles and calcination of raw meal. Complex heat transfer and reactions (solidsolid, gassolid and homogeneous reactions in gas

phase) in rotary kiln were modeled using three sub-models coupled to each other. Solidsolid reactions in the bed region of the kiln were

modeled using pseudo-homogeneous approximation. Melting of solids in the bed and formation of coating within the kiln were accounted.

Clinker cooler was simulated by developing a two-dimensional model to capture cross-flow heat transfer between air and hot clinkers. The

individual models were coupled with each other via mass and energy communication through common boundaries. The coupled model equations

were solved iteratively. The model predictions agree well with the observations and experience from cement industry. The model was used

to gain better understanding of influence of operating conditions on energy consumption in cement plant. Several ways for reducing energy

consumption were computationally investigated. The integrated model, the developed software RoCKS (for Rotary Cement Kiln Simulator)

and results presented here will be useful for enhancing our understanding and for enhancing the performance of clinker manufacturing. 2007 Elsevier Ltd. All rights reserved.

Keywords: Cement; Energy consumption; Reaction engineering model

1. Introduction

Cement making processes are extremely energy consuming.

Typically for producing one ton of cement, a well-equipped

plant consumes nearly 3 GJ. For each ton of clinker produced,

an equivalent amount of green house gases are emitted. The

manufacture of cement has been the focus of considerableattention worldwide because of the high energy usage and high

environmental impact of the process. Considering the recent

impetus on reduction in emission of green house gases and re-

duction in energy consumption, there is a renewed emphasis

on developing computational models for cement industry and

using this understanding for performance enhancement.

Corresponding author. Tel.: +9120 25902170; fax: +9120 25902621.

E-mail address:[email protected](V.V. Ranade).

0009-2509/$- see front matter 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ces.2007.01.063

A schematic of typical clinker making process is shown in

Fig. 1. The raw meal consisting of predetermined quantities

of CaCO3, SiO2, Al2O3 and Fe2O3 are passed sequentially

through pre-heater, calciner, kiln and cooler to form cement

clinkers. In a pre-heater section the raw meal is pre-heated to

calcination temperature via hot gases coming from calciner. In

a calciner, raw meal is partially calcined. The energy requiredfor endothermic calcination reaction is provided by combusting

a suitable fuel. In most cases, coal is used to provide the re-

quired energy, especially in India. The calciner is supplied with

tertiary air from the cooler and air coming out of kiln exhaust.

The former is to supply sufficient O2 for coal combustion and

later to utilize the heat of kiln gases to enhance calcination

reaction. The hot gases from calciner are sent to pre-heater as-

sembly for pre-heating the solids. The partially calcined solids

from the calciner are fed slowly to a rotary kiln. In the rotary

kiln, remaining calcination and other clinkerization reactions
http://www.elsevier.com/locate/cesmailto:[email protected]:[email protected]://www.elsevier.com/locate/ces


2/18

K.S. Mujumdar et al. / Chemical Engineering Science 62 (2007) 25902607 2591

Cooled clinker

Calcineous

Raw meal

Pre-heater

Assembly

Calciner

Kiln

Cooler

Tertiary Air

Secondary Air

Coal

Air to cooler

Kiln Exhaust

Hot gases

to pre-heater

Pre-heated

Raw meal

Exhaust to

atmosphere

Coal

Vent Air

Fig. 1. Schematic of cement clinker process.

occur (formation of C2S, C3A, C4AF). The energy required for

endothermic clinker reactions is provided by combusting coal

in the kiln. The pulverized coal along with the pre-heated air

(secondary air) is fed to the kiln in a counter current modewith respect to solids. Part of the solids melts in the kiln.

The melt formation causes an internal coating on kiln refrac-

tories. Counter current flow of gas entrains solid particles in

the free board region. Such entrainment enhances rates of ra-

diative heat transfer by increasing effective emissivity and con-

ductivity. The hot clinkers are discharged from kiln to clinker

cooler and hot gases from kiln exhaust are sent to the cal-

ciner. In a clinker cooler, a part of energy of solids is recov-

ered back by heat exchange with air. The pre-heated air from

the coolers is passed to kiln and calciner as secondary and

tertiary air, respectively. A small part of air may be vented if

required.This brief overview of clinker formation clearly demonstrates

the strong coupling among pre-heater, calciner, kiln and cooler.

It is therefore essential to develop an integrated model for pre-

heater, calciner, kiln and cooler in order to capture key char-

acteristics of clinker manufacturing and to enable the model to

be used as simulation or optimization tool. Such an attempt is

made in this work.

Recently some attempts have been made to develop

computational fluid dynamics (CFD) based models to simulate

either calciner (for example,Lu et al., 2004) or kiln (for exam-

ple, Mastorakos et al., 1999; Mujumdar and Ranade , 2003).

Though such CFD models show promise in simulating details

of combustion and burner designs, it is almost impossible to

build CFD models for simultaneous and coupled simulations

of pre-heaters, calciner, kiln and cooler. The CFD models are

thus not very useful to gain understanding of coupling and

exploring ways to reduce overall energy consumption per tonof clinker. Some attempts have also been made to develop

reactionengineering models for kiln (for example, Mujumdar

and Ranade, 2006; Spang, 1972). Such models have shown

promising capabilities in capturing the overall behavior and

providing useful clues for reducing energy consumption in

rotary cement kilns. The numerical experiments using the com-

putational model could also predict the influence of kiln oper-

ating parameters on net energy consumption (NEC) in kilns.

Such guidelines can provide useful hints to operating engineers

for kiln optimization. However, none of these models have in-

cluded coupling of pre-heater, calciner, kiln and clinker cooler.

This work was undertaken to fulfill this need. The motivationof the present work was to develop a framework of reaction

engineering based computational model for clinker formation

in cement industry and use this framework subsequently for

exploring possible performance enhancement. The paper is

organized as follows.

The key issues in modeling individual models are discussed

in Section 2. The computational model and the modeling strat-

egy are thereafter presented in Section 3. Section 4 reports

the results of computational simulations of model with respect

to key operating parameters. The use of the developed model

to explore possible ways of reducing energy consumption in

kiln is discussed in Section 5. Key findings of the study are

summarized at the end.
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


3/18


4/18


However, all the attempts for prediction of residence time in

cyclones were based on lab scale cyclones and none of the

studies were extended or reported for industrial scale cyclones.

This parameter was therefore treated as an adjustable param-

eter in the model. In the present work, we have adjusted the

residence time so as to get desired degree of calcination as per

industrial observations. It was confirmed from our prior simula-tions(Warudkar et al., 2005) that varying residence time in the

calciner by 10% had relatively small effect (2.5%) on pre-

dictions of percentage calcination. It is also essential to obtain

relevant kinetics for calcination reaction in calciner. Thermal

decomposition of limestone calcination is a complex process.

A wide discrepancy is observed in the proposed rates for cal-

cination reaction. In our recent work (Mujumdar and Ranade,

2006) we have compared models proposed by 18 investigators

which showed wide scatter. Watkinson and Brimacombe (1982)

have reported experimental data on calcination of limestone in

experimental kiln. The experimental conditions of their exper-

iments were close to industrial operations (bed temperature

1000.

1300 K). Their data was therefore used to find calcination

kinetics in this work.

2.3. Rotary kiln

The partially calcined raw meal is passed slowly to the rotary

kiln where the clinkerization reactions occur. In the initial part

of the kiln the remaining calcination occurs. Other solidsolid

and solidliquid clinkerization reactions take place as the solid

bed moves towards the burner. Part of the solids melts in the

kiln. The melt formation causes an internal coating on kiln re-

fractories. Counter current flow of gas entrains solid particles

in the freeboard region. Such entrainment enhances rates of ra-diative heat transfer by increasing effective emissivity and con-

ductivity. In this section we discuss the key issues involved in

modeling the cement kilns very briefly. The main key issues for

modeling the rotary cement kilns are estimating the residence

time of solids in the kiln, cinkerization reaction in bed region,

coal combustion in freeboard region, heat transfer between bed

freeboard and walls, melting/coating formation around the kiln

walls. These issues are discussed in detail in our recent work

(Mujumdar et al., 2006)and therefore are not repeated here.

2.4. Clinker cooler

The hot solids from the kiln are discharged on the grate of

clinker cooler. As the grate moves with uniform speed along

the cooler length, solids lose their heat to cross-flow air. A part

of the air is generally sent to the kiln as secondary air, a part to

calciner as tertiary air and a part is vented to the surroundings

(vent air). The most important key issue in modeling grate cool-

ers is predicting the heat transfer coefficient between hot solids

and cross-flow air. There is no information on modeling of heat

transfer in such cases. In absence of any relevant information

we have used heat transfer correlation in packed bed reactors

to estimate the heat transfer. Nsofor and Adebiyi (2001) have

carried experimental measurements and presented correlation

for forced convection gas particle heat transfer coefficient for

wide range of temperatures (2001000 C). Since the temper-

atures in clinker cooler are in the same range this correlation

was used to model heat transfer coefficient between solids and

gas. The computational models for individual components and

the coupling strategy are discussed in the following section.

3. Computational models and solution methodology

3.1. Cyclone pre-heater model

A schematic of pre-heater unit considered for developing

computational model is shown in Fig. 3a. The present frame-

work of computational models was developed for dry process

of clinker formation since this process is widely used in ce-

ment industry. For the dry processes, the moisture content is

generally present in very small amount (typically 0.5%, see

for example Engin and Ari, 2005; Peray, 1984). The energy

requirements for removing the moisture from the feed being

small (less than 0.5% of the total energy consumption), the feed

was considered to be free of moisture in this work. However,

the developed framework is quite general and including evap-

oration of moisture from the feed is straightforward. The gas

phase and solids in a cyclone was assumed to be completely

back mixed. In Fig. 3a, Ms is the mass of solids entering the

cyclone.Mg is the mass of the air entering the cyclone. Mse is

the mass of solids entrained from a cyclone. Each cyclone was

assumed to be lined with refractory of thicknesstr .

Thus, for any i th cyclone in pre-heater assembly the follow-

ing inlet streams were considered:

1. Solids from the (i 1)th cyclone (Ms,i1 at temperature

Ti1).2. Solids that are entrained by gas from (i +1)th cyclone

(Mse,i+1 at temperature Ti+1).

3. Air from(i +1)th cyclone (Mg at temperature Ti+1).

The outlet streams for this cyclone are:

1. Solids going out of cyclone (Ms,i at temperature Ti ).

2. Solids that are entrained by gas (Mse,i at temperature Ti ).

3. Air going out (Mg at temperature Ti+1).

The steady state material balance equation for ith cyclone is

written as

Ms,i1+ Mse,i+1= Ms,i + Mse,i , (1)

Mse,i =(1m,p)Ms,i . (2)

In the above equationsm,p represents the particle capture ef-

ficiency of thei th cyclone.Mrepresents the mass of the solids

(in kg/s) and subscripts s andse represent solids and entrained

solids, respectively, as explained earlier.

The steady state energy balance for the i th cyclone was writ-

ten as

Ms,i1 Cp,s Tc,i1+ Mse,i+1 Cp,s Tc,i+1

+Mg Cp,g Tc,i+1

=Ms,i Cp,s Tc,i + Mse,i Cp,s Tc,i

+ Mg Cp,g Tc,i +hcyc Acyi (Tc,i Tiw,i ). (3a)
http://-/?-http://-/?-http://-/?-


5/18

2594 K.S. Mujumdar et al. / Chemical Engineering Science 62 (2007) 2590 2607

i

Tiw,i

1. Refractory

2. Shell

Mg,Mse,i-1Ti-1

Mg,Mse,iTi

Ms,i,Ti

12

Tow,i

Radiation and convection

Losses

Hot clinker, Ts, in Cold clinker, Ts, out

Secondary air Tertiary air Vent air

Cooling Air, Ta

T=0

x

T=0

y

x

T=0

T=0

y

L

Loses

Air in

Gas out

Coal in

Partially calcined raw meal

Raw meal

Ms,i-1Ti-1

Fig. 3. (a) Schematic of (a) cement pre-heater, (b) cement calciner, and (c) grate cooler.

In the aboveCp,s and Cp,g represents the specific heat of solids

and air, respectively. Subscript g represents the air and Tc,irepresents the temperature of solids and air in the i th cyclone.

hcycrepresents the heat transfer coefficient for energy exchange

between particle laden gas and cyclone inner walls. hcyc was

evaluated from the following empirical correlation given by

Gupta and Nag (2000)for heat transfer in cyclones:

hcycdc

kg=702.818+9.02871014u0Re+11.1385

Pu20

+ 4.50398105P

u0

Re+Rc,

where

Rc = Fpw

T4iw T

4g

Tiw Tg

dc

kg. (3b)

The LHS of Eq. (3a) thus represents the total energy entering

the cyclone and RHS represents the energy leaving out of the

cyclone. At steady state the heat given to cyclone walls must

be same as heat conduction in through refractory and cyclone

walls, which is equal to loss from shell walls due to convection

and radiation. The energy balance for heat transfer in cyclone

cross-section is written as

hcyc Acyi [Tc,i Tiw,i ] =2Lkr [Tiw,i Tr,i]

ln(rr /ri ), (4)

2Lkr [Tiw,i Tr,i]

ln(rr /ri )

=2Lksh [Tr,iTow,i ]

ln(r0/rr )

,

(5)

2Lksh [Tr,i Tow,i ]

ln(r0/rr )

=hconv Acyo [Tow,i T0] + cy Acyo [T4

ow,i T4

0].

(6)

In the above equations,Tiw,i is the internal wall temperature of

thei th cyclone,Tr,iis the temperature of interface of refractory

and shell, Tow,i is the temperature of external wall of the ith

cyclone andT0is the ambient temperature.L is the total height

the cyclone, kr is the thermal conductivity of the refractory

and ksh is the thermal conductivity of cyclone walls. r0 is the


6/18


7/18


The individual species balances for limestone and calcium

oxide are given as

dmCaCO3,L

d = rc, (18)

dmCaO,L

d =rc MwCaO

Mw CaCO3. (19)

The energy balance for the raw meal particle is given as

d(mp,LCp,s TL)

d =hc,L

LAp (Tg TL)

+ LLAp (T

4g T

4L )

+ Cp,s TLdmp,L

d , (20)

where TL is the temperature of raw meal particle, L is the

emissivity of solid particle, LAp is the area of the raw meal

particle (which was calculated based on conversion i.e., mass ofraw meal particle reacted) and is residence time of raw meal

particle in the calciner.hc,L was estimated by using Eq. (11).

3.2.2. Continuous phase

The over all gas mass balance is given as

dmg

dt=mgin mgout+ [mp,cin mp,cout]

c Np

+ [mp,Lin mp,Lout ]LNp, (21)

where mg is the mass of the air in the calciner, cNp is the

number of particles of coal coming in per unit time and L

Npis the number of particles of raw meal coming in per unit time.

The individual species mass balance for rate of change of mass

of oxygen, carbon-dioxide, volatile matters and water can be

written as

dyO2dt

=1

mg

mg,in yO2in mg,out yO2out

[mp,c yc,cin mp,c yc,cout ]

c Np MwO2

Mw char

rcombg

Mw vol Vreact MwO2 ZO2 yO2 dmg

dt ,

(22)

dyCO2dt

=1

mg

mgin yCO2in mgout yCO2out

+[mp,c yc,cin mp,c yc,cout]

c Np M wCO2

Mwchar

+ [mp,Lin mp,Lout ]LNp

+

rcombg

Mw vol

Vreact M wCO2 ZCO2

yCO2

dmg

dt , (23)

dyv

dt=

1

mg

mg,in yv,in mg,out yv,out

+ [mp,cin yv mp,cout yv] c Np

rcombg

Mwvol Vreact M wvol Zvolyvol

dmg

dt

, (24)

dyw

dt=

1

mg

mgin ywin mgout ywout

+

rcombg

Mw vol

Vreact Mww ZH2O yw

dmg

dt

,

(25)

where yO2, yCO2, yv, yw are the respective mass fractions of

oxygen, carbon-dioxide, volatile matters and water. Mw O2,Mw CO2 , Mwvol and Mww are their respective molecular

weights.Vreact is the volume of reactor. Subscripts in and outrepresent the inlet and outlet conditions and Z is the stoichio-

metric coefficient.

The energy balance equation for the gas phase is given as

dmgCp,g Tg

dt=mgin Cp,g Tg,in mgout Cp,g Tg

+Sgcomb+ Sccomb+ Scalc

hcyc Acyi (Tg Tiw ), (26)

where

Sgcomb= rcombg Hcombg Vreact, (26a)

Sccomb=

0

hc Ap (Tg Tcl)+c Ap (T

4gT

4cl)

+ Cp,c Tcldmp,c

d

(1fc)Hcomb rcomb

d, (26b)

Scalc=

0

hc,L

LAp (Tg TL)+LLAp (T

4g T

4L )

+ Cp,s TLdmp,L

d +rc Hcalc d. (26c)

In the above equations, Sgcomb, Sccombare the heat source term

for gas-phase from volatile combustion and char combustion,

respectively. Scalcis the heat sink term from calcination. Hcombg,

Hcalcare the enthalpies of volatile combustion and calcination.

hc,Lis the convective heat transfer coefficient between raw meal

particles and air. The steady state equations across the cyclone

walls were written same as that of pre-heaters explained in

the previous section to obtain temperature of calciner internal

walls, refractory and outer walls.

The calciner model equations were solved using an iterative

method. The model equations for gas phase were solved as-

suming steady state. For the first iteration, source terms from

discrete phase were assumed to be zero. The temperature and


8/18


mass of species obtained by solving continuous phase were used

in discrete phase equations to get the new source terms from

the discrete phase. The sources from the discrete phases were

passed to continuous phase to get the new mass and temperature

terms for discrete phase. This procedure was continued till the

subsequent changes in temperature of gas phase were within

0.1%. Suitable under-relaxation parameters were defined toaccelerate convergence. Typically about 20 iterations were re-

quired to achieve convergence. The differential equations for

discrete phase were solved by modified Gears method imple-

mented in ODEPACK (Hindmarsh, 1983). The algebraic equa-

tions for continuous phase were solved using NewtonRaphson

method.

3.3. Kiln model

A comprehensive one-dimensional model was developed to

simulate complex processes occurring in rotary cement kilns.

A modeling strategy comprising three sub-models viz. model

for simulating variation of bed height in the kiln, model for

simulating clinkerization reactions and heat transfer in the bed

region and model for simulating coal combustion and heat trans-

fer in the freeboard region was developed. The Kramers model

(Kramers and Croockewit, 1952)which relates volumetric flow

rate of solids,v, with kiln tilt (, radian), angle of repose (,

radian), radius of kiln(R,m), rotational speed of kiln (n) and

height of solids (h) was used to model bed height variation in

the kiln. The clinkerization reactions in solid bed were modeled

assuming solids as pseudo fluids. Melting of solids in bed re-

gion and formation of coating within the kiln were accounted.

Combustion of coal in the freeboard region was modeled by

accounting devolatilization, finite rate gas phase combustionand char reaction. Knowing the bed and freeboard gas temper-

atures, the temperatures of kiln inner wall, refractory and shell

were obtained by solving steady state energy balance across

the kiln walls. The details of the models and model equation

are discussed in detail in our recent publication (Mujumdar

et al., 2006) and are not repeated here for the sake of brevity.

3.4. Cooler model

The mathematical model of cooler was based on a schematic

shown inFig. 3c. Solids of uniform particle size and constant

porosity were assumed to move in a plug flow with constantgrate speed. Air was assumed to enter in a cross-flow mode

with respect to solids in y direction. The amount of air fed to

the cooler was distributed as secondary air (to kiln) from the

front section of cooler, followed by the tertiary air (to calciner)

and finally the vent air (Locher, 2002) as shown in Fig. 3c.

The amount of secondary, tertiary and vent air (in kg/s) go-

ing to kiln, calciner and exhaust, respectively, were assumed

to be proportional to the fraction of length of each section in

the cooler. The fractional length of each section was user in-

put to the model. To get the temperature profiles of solid bed

and air, the clinker cooler was divided into n segments along

the length of the cooler and m segments along the height of

the cooler. Mass and energy balances were solved for these

segments. Conductive heat transfer was considered for solids in

both horizontal and vertical directions. Convective heat transfer

coefficient between air and solids was calculated from empiri-

cal correlation assuming solids as packed bed as discussed pre-

viously. The boundary conditions used in the model are shown

inFig. 3b. The model equations are presented in the following.

Mass balance for solids can be written asdms(i,j)

dx=0. (27)

Assuming steady state operation, the energy balance equation

can be written as

(s (1)us,x Cp,s Ts )

x+

(s (1)us,y Cp,s Ts )

y

={(1)ksTs /x}

x+

{(1)ksTs /y}

y

a hc,c (Ts Tg). (28)

In this equations

is the cement clinker density,Cp,s is clinker

heat capacity, us is grate speed, and Ts is clinker temperature

of solid at any point,ks is clinker thermal conductivity,a is the

surface area per unit volume, hc,c is convective heat transfer

coefficient between solid clinker and air in the cooler, is

the porosity, Tg is air temperature at any point in the cooler.

In Eq. (28) the first and second terms of the right-hand side

represents the conductive heat transfer. The last term in right-

hand side represents convective heat transfer between the air

and solids.

The mass balance for air can be written as

dma(i,j)

dy=0. (29)

Energy balance for air can be written as

(gug,x Cp,g Tg)

x+

(gug,y Cp,g Tg)

y

={kgTg/x}

x+

{kgTg/y}

y+a hc,c (Ts Tg).

(30)

In this equation g is the density of the air,ug,y is inlet speed

of cooling air, and Tg is air temperature at any point, k is air

thermal conductivity,Ts is solid temperature at any point in the

cooler. In Eq. (30) the left-hand side terms represents the net

energy input by the air. First two terms in the right-hand side

represent the conduction between the air layers and the final

term is due to the convection between solids and air.

In Eq. (30) hc,c is convective heat transfer coefficient be-

tween solid clinkers and air. Developing accurate models for

convective heat transfer coefficient between solids and air is

important in capturing heat transfer in the cooler. In this work

the convective heat transfer coefficient was calculated based on

empirical expression given by Nsofor and Adebiyi (2001). The

empirical expression is given as

Nu=8.74+9.34 [6(1)]0.2 Re0.2 Pr0.33

(30


9/18


It is important to note that the Reynolds numbers for commer-

cial clinker cooler are significantly higher (Re 1000.2000)

as compared to the experimental conditions of Nsofor and

Adebiyi (2001). However, as discussed earlier, there are no

other systematic experimental studies reported to predict con-

vective heat transfer coefficients in clinker coolers. The empiri-

cal correlation (Eq. (31)) was developed for particle sizes closeto those found in industrial clinker coolers and for wide range

of temperature conditions as observed in clinker coolers. Fortu-

nately, the empirical correlation seems to be weekly dependent

on Reynolds number (Reynolds number is to power 0.2). There-

fore possible errors associated with Eq. (31) are not expected

to change the simulation results significantly (predicted Nus-

selt number is 50.60). Hence Eq. (31) was used to predict

gas solid heat transfer in clinker coolers in the present model.

All the physical properties for determining heat transfer coef-

ficient were calculated at an average temperature of solids and

air as Tf =(Ts +Tg)/2. The system of algebraic linear equa-

tions formulated for above model equations was solved using

tri-diagonal matrix algorithm (TDMA).

3.5. Integrated model and solution strategy

The individual models for pre-heater, calciner, kiln and

cooler described in the previous section were coupled with

each other to develop a simulator for the entire system. The

schematic of the simulator is shown inFig. 4.The required in-

puts to the simulator are flow rates and composition of (a) raw

meal entering the pre-heater, (b) air entering the cooler, (c) coal

entering the calciner and the kiln, and (d) the material prop-

erties and operating parameters of the individual equipments

(for example, kiln RPM, grate speed of cooler). However, tosolve the integrated simulator, it is necessary to know the in-

let conditions for the calciner (flow rate, mass fractions and

temperature of solids and air from pre-heater, kiln and cooler),

pre-heater (flow rate and temperature of air from calciner), kiln

(flow rate, mass fractions and temperature of secondary air

from cooler and partially calcined raw meal from the calciner)

and cooler (flow rate and temperature of solids from kiln).

To generate these inputs a pre-processor was developed. The

function of pre-processor was two-fold. The pre-processor was

used to develop good initial guess for the simulator and also to

check for any inconsistency of input data. The pre-processor

generated the initial guess (for mass flow-rates, composition

and temperatures of raw meal and air) for the individual mod-els based on overall material and energy balances. Following

parameters were provided to the pre-processor to achieve this:

1. Percentage calcination occurring in the calciner(P ).

2. Temperature of secondary air (Tg,S)and tertiary air (Tg,T)

leaving the cooler.

3. Temperature of air leaving the kiln (Tg,K ).

4. Temperature of air exiting the pre-heater to the atmosphere

(Tg,P).

5. Temperature of solids exiting the cooler (Ts,R ).

6. Heat losses (HLoss,K ) and heat of clinkerization reaction in

the kiln (HR,K ).

These values are usually known or can be easily available for

any cement plant and can therefore be used to generate good ini-

tial guess for faster convergence of solution. The pre-processor

solves mass and energy balance equations as discussed in the

following. Based on the percentage calcination in the calciner,

the mass of CO2 produced in calciner was calculated as

mCO2,C = mCaCO3,i P MwCO2

Mw CaCO3, (32)

where mCaCO3,i is the total amount of CaCO3 in the inlet raw

meal. The mass flow rate of solids entering the kiln was calcu-

lated as

Ms,C =Ms,P mCO2,C , (33)

where Ms,C is the mass flow rate of raw meal leaving the

calciner or entering the kiln, Ms,Pis the mass flow rate of the

solids entering the pre-heater. The corresponding mass fraction

of solids species leaving the calciner or entering kiln were

calculated as

xCaCO3,C =mCaCO3,i mCaCO3,i P

Ms,C,

xCaO,C =(mCO2,C )(M wCaO)

(Ms,C )(MwCO2 ),

xSiO2,C =mSiO2,i

Ms,C, xAl2O3,C =

mAl2O3,i

Ms,C,

xFe2O3,C =mFe2O3,i

Ms,C, (34)

wherex is the mass fraction of the component in the raw meal.The amount of clinker leaving the kiln or entering the cooler

Ms,K was calculated as

Ms,K =Ms,C (Ms,C )xCaCO3,C

MWCaCO3MwCO2 . (35)

Based on overall material balance on kiln, the amount of air

leaving the kiln was calculated as

Mg,K =Mg,S+Ms,C xCaCO3,C Mw CO2

Mw CaCO3

+ Mc,K

yc,K

MwCO2

MwCaCO3, (36)

where Mg,Sis the mass of secondary air entering the kiln, Mg,Kis the air leaving the kiln or entering the calciner, Mc,K is the

amount of coal entering the kiln and yc,K is the mass fraction

of char entering the kiln. The amount of air leaving the pre-

heater assembly was calculated as

Mg,P =Mg,K +Mg,T +mCO2,C

+ Mc,C yc,c Mw CO2

Mw CaCO3, (37)

where Mg,P is the mass of air entering the pre-heater, Mg,T

is the mass tertiary air entering the calciner, mCO2,C is the


10/18


Call PreprocessorConsistence Checks

&

Generate initial guess

Call Sub-models

Update variables

Converged

YN

No

Yes

Post Processing

User Input

Dimensions, MOC,

Mass Flow Rate,

Mass Fractions,

Temperature

Fig. 4. Solution methodology of the simulator.

CO2produced in calciner due to calcination reaction andMc,Cis the amount of coal entering the calciner and yc,c is the mass

fraction of char entering the calciner. The temperature of solids

leaving the kiln was calculated as

Ts,K =(Ms,R Cp,s Ts,R + Mg,T Cp,g Tg,T +Mg,S Cp,g Tg,S)(Mg,in Cp,g Tg,in)

(Ms,K Cp,s ). (38)

In the above equation, Mg,in and Tg,in are the mass flow rate

and temperature of air entering the cooler and Ts,R is the tem-

perature of solids exiting the cooler. The temperature of solids

entering the kiln or exiting calciner (Ts,C ) is calculated as

Ts,C =(Ms,K Cp,s Ts,K +Mg,K Cp,g Tg,K +HR,K +HLoss,K Mg,S Cp,g Tg,SHc,K )

(Ms,C Cp,s ). (39)

In the above equation, Hc,K is the heat released due to coal

combustion in kiln,HR,K is heat required for clinker reactions

andHLoss,K is the loss from the kiln. The temperature of solids

entering the kiln is essentially same as temperature of gases

leaving the calciner (Tg,C ). Finally, the temperature of solids

entering the calciner or leaving the pre-heater assembly (Ts,P)

was calculated as

Ts,P =(Ms,C Cp,s Ts,C +Mg,C Cp,g Tg,C + Hcalc Mg,K Cp,g Tg,K Mg,T Cp,g Tg,T Hc,C )

(Ms,P Cp,s ). (40)

In the above equation, Hc,Cis the heat released due to coal com-

bustion in the calciner and Hcalc is the heat required by calci-

nation reaction. This was easily calculated based on percentage

calcination occurring in the calciner. The heat losses in calciner

are negligible as compared to total heat supplied to the calciner

(< 5% of total energy input) and therefore was not considered

in pre-processor calculations. In this way the input conditions

(mass, mass fractions and temperature) for pre-heater, calciner,

kiln and cooler were calculated using pre-processor. The values

calculated by pre-processor were passed as input conditions to

the individual models. The individual models were then solved

iteratively as shown inFig. 4.The iterations were continued till

the temperature of solids and gases at exit of individual com-

ponents were within error of 1%. Suitable under-relaxation

parameters were used. Typically 1020 iterations were required

for solution to converge. We have also carried out several test

simulations of limiting cases to verify that implemented nu-

merical techniques and computer programs are correctly solv-

ing the model equations. For example, the calciner and kiln

were solved by switching off the calcination and clinkeriza-

tion reactions in the calciner and kiln, respectively. For these

simulations the material and energy balances converged to an

error of 1% giving verification that numerical calculations

are correctly solving the model equations. It was also verified

that the converged solution is not a function of initial guess or

under-relaxation parameters. An easy to use, graphical user

interface (GUI) based software called RoCKS (Rotary Cement

Kiln Simulator) was developed based on the integrated modules

of pre-heater, calciner, kiln and cooler.

4. Results and discussion

The integrated model (RoCKS) presented in the previous

section was used to simulate performances of pre-heater, cal-

ciner, rotary kiln and cooler in clinker manufacturing. Based on

the available data on rotary kilns (Mujumdar et al., 2006) and


11/18


available information from some of the cement industries, a

typical clinker manufacturing configuration was selected as a

base case. Some assumptions were made to fill in the gaps in the

available data. The details of selected configuration are given in

Tables 2ac. Though the developed mathematical framework is

general enough to accommodate temperature dependent phys-

ical properties like heat capacity, at this stage, these propertieswere treated as constants. The physical properties of solids and

air used in this work are specified in Table 3a. Our prior simu-

lations of kiln and calciner (Mujumdar et al., 2006; Warudkar

et al., 2005) indicated that the errors in overall energy

consumption associated with the assumption of temperature

independent values of specific heat were within 1%. The oper-

ation of the base case (described inTables 24)was computa-

tionally studied to understand the various processes occurring

in individual units in clinker formation. On obtaining satisfac-

tory results from the base case, several numerical experiments

were performed using the model for understanding interac-

tions among different processes and for possible optimization

of clinker manufacturing process.

4.1. Base case simulation

The predicted results from the simulation of the base case

are summarized in Table 4. The mass fractions and tempera-

tures of solids and air in pre-heaters, calciner, kiln and cooler

obtained from the simulation are plotted inFigs. 5 and 6, re-

spectively. It is important to note that the flow of air is counter

current with respect to the flow of solids in the system. The ab-

scissas ofFigs. 5 and 6denote particular equipment in clinker

formation as discussed below. Abscissas 14 corresponds to

Table 2

The dimensions of (a) pre-heater unit, (b) kiln and (c) cooler

S/No. Description Units Values

(a)

1 No. of pre-heaters 4

2 Height of cylindrical section m 5

3 Height of conical section m 3

4 Diameter of cyclone m 3

5 Diameter of cone tip m 1

6 Refractory thickness m 0.13

7 Shell thickness m 0.03

8 Inlet duct height m 1

9 Inlet duct width m 1

10 Diameter of outlet pipe m 1

(b)

1 Length m 50

2 Inner diameter m 3.4

3 Coating thickness m 0.136

4 Refractory thickness m 0.2

5 Shell thickness m 0.025

(c)

1 Length m 11

2 Width m 1

Table 3

(a) The physical properties of solids and air, (b) particle size and composition

of coal and (c) particle size and composition of raw meal

Description Air Raw meal Coal

(a)

Thermal conductivity, W/m K 0.116 0.5 0.5

Emmisivity 0.4 0.9 0.8Heat capacity, J/kg K 1000 1000 1000

Viscoscity, kg/m s 1e05

Density, kg/m3 1.3 1500 1000

Char calorific value, kcal/kg 5600

Volatile calorific value, kcal/kg 11 900

(b)

Coal particle size, 50 m

Volatile (CH4)a, 27%

Char, 58%

Ash, 15%

(c)

Raw meal particle size, 50 m

CaCO3, 80%

CaO, 0%

SiO2, 14%

Fe2O3, 3%

AL2O3, 3%

a Mujumdar et al. (2006).

pre-heater assembly. Abscissas 4 and 5 denote the calciner in

the system. Abscissas 515 denote the rotary kiln and 1518

denote the cooler section. Fig. 5 shows a plot of mass frac-

tions in pre-heater, calciner, kiln and cooler (only CaO, C2S

and CO2 mass fractions are plotted for the sake of brevity).Since there is no reaction occurring in pre-heater section, the

composition of CaO and CO2in this section do not vary. How-

ever, in the pre-heater section the raw meal gets heated from

300 to 1069 K and hot gases from calciner get cooled (from

1224 to 539 K) as can be seen fromFig. 6.As the raw meal

passes through the calciner, it gets partially calcined. There-

fore, CaO concentration increases in the calciner section as

can be seen from Fig. 5. Similarly since CO2 is formed due

to calcination and coal combustion, the mass fraction of CO2increases in the calciner. Coal combustion in the calciner ac-

counts for rise in temperature of both solids and gas in the

calciner (seeFig. 6). Remaining clinkerization reactions occurin kiln. The mass fraction and temperature profiles obtained in

kiln (as shown inFigs. 5 and 6) are similar to previously pub-

lished results(Mujumdar and Ranade, 2006; Mastorakos et al.,

1999). Since there is no reaction occurring in the cooler, mass

fraction of solids in the clinker cooler do not vary. However,

air entering the cooler gets pre-heated (from 300 to 1200 K)

and solids get cooled (from 1632 to 476 K) in the cooler sec-

tion. The predicted energy requirements of individual processes

like clinkerization reactions, losses, melting predicted by the

model are listed inTable 4. The obtained results are qualita-

tively similar to previously published results (Engin and Ari,

2005). The performance of the overall system was characterized

in terms of NEC per unit weight of product (clinker coming out
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


12/18


Table 4

Complete energy balance of the system

S/No. Description Pre-heater Calciner Kiln Cooler

1 Solid inlet temperature, K 300 1069.2 1214.8 1622.4

2 Mass flow rate, kg/s 50 50 37.74 32.3

3 Air inlet temperature, K 1214.8 1114.5 1229.9 300

4 Air flow rate, kg/s 60.8 46.7 16.2 455 Coal flow rate, kg/s 2.15 0.9

6 Coal inlet temperature, K 350 350

7 Heat with solids in, kJ/kg clinker 463.0 1650.2 1415.3 1622.4

8 Heat with air in, kJ/kg clinker 2297.9 1603.6 615.0 416.7

9 Heat with coal in, kJ/kg clinker 23.2 9.7

10 Combustion of coal, kJ/kg clinker 1876.7 747.1

11 Heat of reaction, kJ/kg clinker 1384.5 219.0

12 Heat of melting, kJ/kg clinker 44.2

13 Heat of solids leaving, kJ/kg clinker 1650.2 1415.3 1622.4 463.0

14 Heat of air leaving, kJ/kg clinker 1014.4 2297.9 1603.6 1415.3

15 Heat of vent air in cooler, kJ/kg clinker 109.4

16 Heat with ash, kJ/kg clinker 3.22 2.24

17 Loses, kJ/kg clinker 98.9 43.5 140.7

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Divisions across the cement clinker process

Massfraction

CaO -Solid mass fraction

C2S-Solid mass fraction

CO2-Gas mass fraction

Pre-Heaters

Calciner

Kiln

Cooler

Fig. 5. Solid and gas mass fractions in pre-heaters, calciner, kiln and cooler

in a cement clinker process.

of the kiln). The NEC is calculated as

NEC=(ERXN,C +ERXN,K +EMELT,K )+ ELOSS+(EG,OUT+ ES,OUT EG,IN ES,IN). (41)

In the above equation, ERXN denotes the energy required for

clinkertization reactions and subscripts C and K denotes the

calciner and the kiln, respectively. The term EMELT,K denotes

the energy required for melting in the kiln. ELOSS denotes the

summation of energy losses from pre-heater assembly, calciner

and kiln. The other terms denote energy flow rates (subscripts

IN or OUT) for the gas and solid streams (subscripts G or

S) which denote the energy required to raise the sensible heat

of the solids. Based on above calculations, the NEC predicted

by the integrated simulator, for these operating conditions was

2635 kJ/kg clinker (630 kcal/kg clinker) which seems to be

reasonable when compared with industrial observations. Over-

all the integrated simulator was able to predict the clinker man-

ufacturing process in cement industry reasonably well.

4.2. Influence of key design and operating parameters on NEC

On obtaining a reasonable agreement, the model was used to

explore space of design and operating parameters to understand

influence of these parameters on the performance of clinker

manufacturing. All these simulations were carried for a fixed

product composition (C3S mass fraction 0.48 in the product).This was achieved by altering coal flow rate either to calciner

or kiln. This analysis is presented in the section below.

4.2.1. Effect of number of pre-heaters

The effect of changing number of pre-heaters in pre-heater

assembly (from 3 to 5) on NEC was studied. For this simulation

the coal in the kiln was adjusted to get same product composi-

tion at the kiln exit. The results for this simulation are shown

in Fig. 7. It can be seen from Fig. 7 that as number of pre-

heaters in pre-heater assembly increases, solids get pre-heated

to a higher temperature before they enter the calciner (see sec-

ondary axis in Fig. 7). Therefore the coal requirement for a

fixed product composition decreases. Thus the NEC decreases

as number of pre-heaters increases. However, the overall cap-

ital cost increases by increasing number of pre-heaters in the

system. The developed model will be useful to carry out cost

to benefit analysis for introducing additional pre-heater in the

pre-heater assembly.

4.2.2. Effect of percentage calcination in the calciner

The pre-calcination of raw meal in calciner is an important

process in cement process. We have studied the effect of per-

centage calcination in calciner on NEC. To vary the percent-

age calcination in calciner the coal feed rate to the calciner

and kiln was altered for the same clinker composition. The


13/18


0

500

1000

1500

2000

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Divisons accross the cement clinker process

Gastemperature,

K

0

500

1000

1500

2000

Solidtemperatu

re,

K

Solid temperature

Gas temperature

Pre-heaters

Cooler

Kiln

Calciner

Solids leaving 3rd preheater

Gases entering 3rd preheater

Fig. 6. Temperature profile across pre-heaters, calciner, kiln and cooler in a cement clinker process.

615

620

625

630

635

640

645

2 3 4 5 6

Number of pre-heaters

Energyc

onsumption,

kcal/kgclinker

1000

1020

1040

1060

1080

1100

Solidtemperatureenteringcalciner,K

Fig. 7. Effect of pre-heater number on overall energy consumption.

simulation results are shown in Fig. 8.As can be seen from

Fig. 8, the NEC was found to decrease till 70% calcination and

then it increases with further increase in percentage of calcina-tion. The secondary axis ofFig. 8shows that the kiln exit gas

temperature also shows a similar trend. Table 5shows a com-

plete comparison of heat of reaction occurring in kiln and cal-

ciner in this process. The heat of reaction in kiln decreases as

the percentage calcination increases in calciner. The total heat

of reaction in kiln is the summation of heat of calcination (en-

dothermic reaction) and the heat of clinker formation (exother-

mic reactions). When calcination occurs pre-dominantly in the

pre-calciner (> 70%), the energy requirements for reactions in

kiln reduce drastically. This causes increase in kiln flue gas

temperature and increase in losses from kiln shell. Therefore

the NEC and kiln flue gas temperature increases if more than

70% calcination occurs in the calciner. The model and the

626

631

636

641

646

651

656

661

666

40 50 60 70 80 90 100

% Calcination

EnergyCo

nsumption,

kcal/kgclinker

1000

1100

1200

1300

1400

1500

Kilnex

itgasTemperature,

K

Fig. 8. Effect of percentage calcination on overall energy consumption.

Table 5

Heat of reaction in calciner and kiln

S/No. Heat of reaction

1 Calcination, % 50 60 70 80 90

2 Heat of reaction

in calciner, kJ/kg

clinker

1038.5 1176.6 1384 1618.5 1802.5

3 Heat of reaction in

kiln, kJ/kg clinker

582.9 433.2 219 20.2 178.4

RoCKS software were thus able to provide valuable clues for

determining the optimum percentage calcinations desired for

minimizing NEC.


14/18


615

620

625

630

635

640

0 0.5 1 1.5 2 2.5 3 3.5

Kiln tilt, Degree

Energy

consumption,

kcal/kg

clinker

400

800

1200

1600

2000

Residencetimeofsolids,s

615

620

625

630

635

640

2 3 54 6 7 8 9

Kiln Rpm

EnergyConsumption,

kc

al/kg

clinker

500

1000

1500

2000

2500

Residencetimeofsolid

sinkiln,s

Fig. 9. Effect of (a) kiln RPM and (b) kiln tilt on overall energy consumption.

4.2.3. Effect of kiln RPM, kiln tilt and grate speed of clinker

cooler

The effect of kiln rotational speed and kiln tilt on the overall

performance is shown inFig. 9a and b. For these simulations

the coal flow rate to the kiln was varied to maintain constant

product composition. It can be seen fromFig. 9a as kiln RPMdecreases, the NEC decreases. Changes in kiln RPM changes

the bed height and the residence time of solids in the kiln as

can be seen fromFig. 9a and b (2002.4 s for 3rpm; 1058.2 s for

5.5 rpm and 703.4 s for 8 rpm). Our simulation results indicate

that it seems to be beneficial to operate kilns at lower rpm as

long as adequate mixing of solids is occurring. FromFig. 9b it

can be seen that energy consumptions in kilns operated at lower

tilt is less as compared to kilns at higher tilt. The grate of clinker

cooler is the important parameter that controls the residence

time of solids and subsequently the heat exchange between hot

solids and counter current air in the cooler. We have studied the

influence of varying grate speed on overall energy consumption.

The results for these simulations are shown inFig. 10.It can be

620

625

630

635

640

645

650

0.06 0.08 0.1 0.12 0.14 0.16 0.18

Grate speed, m/s

Energyconsumption,

kcal/kgofclinker

1050

1100

1150

1200

1250

1300

1350

Secondaryairte

mperature,

K

Fig. 10. Effect of cooler grate speed on overall energy consumption.

seen fromFig. 10that the NEC increases with increasing gratespeed. The increase in grate speed reduces residence time of

solids in the cooler. This results in less convective heat transfer

between solids and air as clearly indicated by temperature of

secondary air plotted in Fig. 10. Therefore the simulation results

indicate that it is better to operate grates in the cooler at lower

speed. The simulations presented here provide useful trends of

energy consumption as a function of key operating parameters

in cement clinker process. This result also gives us a scope to

understand the importance of design parameters (kiln tilt) on

plant performance and can be very useful to plant engineers.

4.2.4. Effect of solid loadingThe predicted results in the form of NEC and corresponding

overall losses for different solids flow rates are shown in Fig. 11.

It can be seen that the NEC per unit weight of product decreases

as solids flow rate increases. This is because the net energy loss

from the entire system decreases as the solid flow rate increases

(see Fig. 11). Thus, it is beneficial from the point of view

of energy consumption to operate the units with higher solids

flow rate. Other operational concerns like increase in dusting

and mixing, however, need to be considered while identifying

maximum solids flow rate specifically for cement kilns.

4.2.5. Effect of coal compositionThe effect of varying coal composition to the kiln on NEC

is shown in Fig. 12. From Fig. 12, it can be seen that the

overall energy consumption does not change significantly with

changing coal composition (ash content 9%, 15% and 40%).

For these simulations the coal flow rate to the kiln was ad-

justed so that the same amount of energy is supplied to the

kiln. Therefore the insignificant change in overall energy con-

sumption does not seem to be surprising. However, as the

coal composition changes, the flame characteristics in the kiln

vary. The predicted dimensionless flame length by the simu-

lator for varying coal composition is shown in Fig. 12. The

flame length was calculated by tracking the region in freeboard

where char and volatiles composition in coal go to zero. The


15/18


600

610

620

630

640

650

44 45 46 47 48 49 50 51

Raw meal flow rate, kg/s

Energyconsumption

,kcal/kgclinker

250

270

290

310

330

350

Totallossesinclinkerpr

ocess,

kJ/kgclinker

Fig. 11. Effect of raw meal flow rate on overall energy consumption.

600

610

620

630

640

650

30 40 50 60 70

Char Percentage, %

Ene

rgycomsumption,

kcal/kgofclinker

0.4

0.44

0.48

0.52

0.56

Flamelength,

dimensionless

Fig. 12. Effect of coal composition on overall energy consumption.

dimensionless flame length was calculated as the ratio of pre-

dicted flame length to length of the kiln. It can be seen that

coal with higher ash content tends to have a longer flame as

compared to coal with lower ash content. The flame length is a

complicated function of amount of oxygen, amount of char and

temperature of gas and particle in the freeboard region. Coalswith higher ash content tends to consume oxygen at a slower

rate and therefore result in longer flames. Such simulations can

therefore provide useful information to kiln operators to pre-

dict the flame characteristics for wide variety of coal available

in the market.

4.2.6. Effect of secondary shell

Heat losses to the surrounding from the kiln shell by radia-

tion and convection are a significant source of energy loss in

cement kilns and therefore the overall process. These losses

can be reduced by using a secondary shell. The idea is to cover

the kiln shell with another metallic shell having low surface

590

592

594

596

598

600

15 20 25 30 35 40 45

Air flow rate through secondary shell, kg/s

Energycomsumption,kcal/kgofclinker

1090

1095

1100

1105

1110

1115

1120

Kilnfluegastemperature,

K

Fig. 13. Effect of secondary shell, on overall energy consumption.

emissivity and thermal conductivity (Engin and Ari, 2005).

However, merely covering kiln shell with metallic shell and

insulating it can lead to enormously high shell temperatures.

Hence a practical approach to use secondary shell would be to

feed air through the interstitial space of shell and secondary

shell to recover the energy and still operate kilns under realis-

tic conditions(Mujumdar et al., 2006). The developed RoCKS

frame work was used to explore the possibility of using such

a secondary shell. The losses in kiln reduced from 140 kJ/kg

of clinker to 1.4kJ/kg of clinker on applying a secondary

shell and insulation of dimensions and operating conditions

specified in Mujumdar et al. (2006). The NEC reduces from 2635kJ/kg clinker to 2493 kJ/kg clinker (i.e., 630 kcal/kg

clinker to 596 kcal/kg clinker) by using secondary shell and

passing air of about 30 kg/s through the interstitial space

(Fig. 13). If the air coming out of annular space at 496K

can be utilized within the cement plant (refrigeration, drying

of fly ash and so on), the use of secondary shell appears to

be promising for reducing NEC in the clinker manufacturing

process.

5. Conclusions

A comprehensive model was developed to simulate complexprocesses occurring in pre-heater, calciner, kiln and cooler for

clinker formation in cement industry. The models for pre-heater

and calciner were developed assuming solids and gas to be

completely back mixed. The computational model for the kiln

was developed assuming gas and solids as plug flow. The inte-

grated simulator was converted into simple to use GUI based

software for cement industry, named as RoCKS. RoCKS was

used to simulate performance of pre-heater, calciner, kiln and

cooler for clinker formation. Detailed validation was unfortu-

nately not possible since adequate industrial data could not be

obtained. However, the model predictions agreed reasonably

with industrial observations. RoCKS was used to understand

influence of various design and operating parameters on overall


16/18


performance. Specific conclusions based on this computational

study are:

Including an additional pre-heater reduces NEC. The devel-

oped model can be used to evaluate relative benefits of en-

ergy savings by additional pre-heater and required additional

capital expenses. There is an optimum value for percentage of calcination car-

ried out in calciner with respect to overall energy consump-

tion in clinker manufacture. With the parameters selected in

this work, this optimum value of percentage calcination in

calciner is about 70.

The simulation results indicated that operating kiln with

higher solid loading, lower rpm, lower tilt and lower grate

speed reduces energy consumption per unit production. The

upper limit on solid loading (bed height) and lower limits on

rpm and tilt (mixing and heat transfer) need to be identified

based on other practical issues.

The use of secondary shell appears to be a promising method

to reduce overall energy consumption, if the hot air generates

in such secondary shell (200 C) can be utilized in some

other processes in cement plants.

The model was also able to predict kiln characteristics like

maximum flame temperature and overall flame length for coals

with different compositions. The models and results presented

here will help in developing a better understanding of clinker

manufacturing process and may provide clues for possible

optimization.

Notation

a surface area per unit volume, m2/m3

A0 devolatilization constant

Acyi internal surface area of cyclone, K

Acyo external surface area of cyclone, K

Ap surface area of coal particle, m2

LAp surface area of solid particle, m2

Cp,c specific heat capacity of coal particle, J/kg K

Cp,g specific heat capacity of air, J/kg K

Cp,s specific heat capacity of solids, J/kg K

dc inner diameter of cyclone, mdp radius of particle, m

E1 energy of activation for char combustion, J/mol

E2 energy of activation for calcination, J/mol

fc fraction of heat given to coal particle released

due to coal combustion

fv,0 initial mass fraction of volatiles in coal particle

Fpw view factor

hc heat transfer coefficient between coal particle

and gas, W/m2 K

hc,c heat transfer coefficient between clinker and gas,

W/m2 K

hc,L heat transfer coefficient between solid particle

and gas, W/m2 K

hcyc heat transfer coefficient between particle laden

gas and cyclone inner wall, W/m2 K

Hc,C heat of coal combustion in calciner, J/kg

Hc,K heat of coal combustion in kiln, J/kg

Hcalc heat of calcination reaction in calciner, J/kg

Hcomb heat of char combustion, J/kg

Hcombg heat of volatile gas phase combustion, J/kgHLoss,K heat losses in the kiln, J/kg

HR,K heat required for clinker reactions, J/kg

kg thermal conductivity of air/gas, W/m K

kr thermal conductivity of refractory, W/m K

ks thermal conductivity of clinker, W/m K

ksh thermal conductivity of shell, W/m K

ks,c rate constant of char combustion, kg/m2 skPa

r ks rate constant of calcination of calcium carbon-

ate, mol/m2 s1

r k

s rate constant of calcination of calcium carbon-

ate, mol/m2 s1

L total height of cyclone, m

ma mass of air in cooler, kg/s

mAl2O3,i mass of total aluminum oxide in solids in cal-

ciner, kg/s

mCO2,C mass of carbon-dioxide produced in calciner due

to calcination, kg/s

mCaCO3,i mass of total calcium carbonate in solids in cal-

ciner, kg/s

mFe2O3,i mass of total ferrous oxide in solids in cal-

ciner, kg/s

mSiO2,i mass of total silicon dioxide in solids in cal-

ciner, kg/s

mg mass of gas in calciner, kg

mgin mass of air entering in calciner,kg/smgout mass of air leaving calciner, kg/s

mg,K mass of air leaving the kiln calciner, kg

mp,c mass of coal particle, kg

mp,cin mass of coal particle entering calciner, kg

mp,cout mass of coal particle leaving calciner, kg

mpc,0 initial mass of coal particle, kg/s

mp,L mass of solid particle, kg

mp,Lin mass of solids entering calciner, kg

mp,Lout mass of solids leaving calciner, kg

ms mass of solids/clinker in cooler, kg

Mc,C mass of coal entering the calciner, kg/s

Mc,K mass of coal entering the kiln, kg/sMg mass of gas in cyclones, kg/s

Mg,K mass flow rate of secondary air entering the

kiln, kg/s

Mg,P mass flow rate of gas entering the pre-heater,

kg/s

Mg,S mass flow rate of secondary air entering the

kiln, kg/s

Mg,T mass flow rate of tertiary air entering the cal-

ciner, kg/s

Ms mass of solids in cyclones, kg/s

Mse mass of solids entrained by gas in cyclones, kg/s

Ms,C mass flow rate of solids leaving the calciner, kg/s

Ms,K mass flow rate of clinker leaving the kiln, kg/s


17/18


Ms,P mass flow rate of solids entering the pre-heater,

kg/s

Mw CaCO3 molecular weight of calcium carbonate, kg/kmol

Mw CaO molecular weight of calcium oxide, kg/kmol

Mw char molecular weight of carbon, kg/kmol

Mw CO2 molecular weight of carbon-dioxide, kg/kmol

Mw O2 molecular weight of oxygen, kg/kmolMw vol molecular weight of volatile, kg/kmol

Mw w molecular weight of water, kg/kmol

Nu Nusselt numbercNp number of coal particles entering calciner per

secondLNp number of solid particles entering calciner per

second

pO2 partial pressure of oxygen in gas, kPa

pCO2 partial pressure of carbon-dioxide in gas, kPa

peq equilibrium partial pressure for carbon-dioxide

in gas, kPa

P the percentage calcination occurring inside the

calciner

P pressure drop across the cyclone, mm of H2O

Pr Prandtl number

rc rate of calcination, kg/s

rcomb rate of combustion of char particles, kg/s

rcombg rate of combustion of volatiles, kg/s

ri internal diameter of cyclone, m

r0 external diameter of cyclone, m

rp radius of solid particle, m

rr internal diameter of cyclone shell, m

R gas constant

Rc non-dimensional form of radiative heat transfer

coefficientRe Reynolds number

T0 ambient air temperature, K

Tc,i temperature of solids and gas in cyclone, K

Tcl temperature of coal particle, K

Tf average temperature of solids and air in

cooler, K

Tg temperature of gas, K

Tg,in temperature of gas entering calciner, K

Tg,out temperature of gas exiting calciner, K

Tg,K temperature of gas leaving the kiln, K

Tg,S temperature of secondary air, K

Tg,T

temperature of tertiary air, K

Tg,P temperature of gas leavingthe pre-heater, K

Tiw,i the internal wall temperature of the cyclone, K

Tow,i the external wall temperature of the cyclone, K

TL temperature of solid particle in calciner, K

Tr,i the temperature of interface of refractory and

shell in cyclone, K

Ts temperature of solids/clinker in cooler, K

Ts,C temperature of solids entering the kiln, K

Ts,R temperature of solids exiting the cooler, K

u0 inlet gas velocity in cyclone, m/s

us,x grate speed in x direction, m/s

us,y grate speed in y direction, m/s

ug,x air velocity inx direction, m/s

ug,y air velocity iny direction, m/s

Vreact volume of reactor, m3

xAl2O3,C mass fraction of aluminum oxide entering kiln

xCaCO3,C mass fraction of calcium carbonate entering kiln

xCaO,C mass fraction of calcium oxide entering kiln

xFe2O3,C mass fraction of ferrous oxide entering kiln

xSiO2,C mass fraction of silicon dioxide entering kilnyc,c mass fraction of char in coal particle in calciner

yc,cin mass fraction of char entering in coal particle

yc,cout mass fraction of char leaving in coal particle

yc,K mass fraction of char in coal particle entering

the calciner

yv,c mass fraction of volatiles in coal particle

yO2 mass fraction of oxygen in gas

yO2,in mass fraction of oxygen entering calciner in gas

yO2,out mass fraction of oxygen leaving calciner in gas

yCO2 mass fraction of carbon-dioxide in gas

yCO2,in mass fraction of carbon-dioxide entering cal-

ciner in gas

yCO2,out mass fraction of carbon-dioxide leaving calciner

in gas

yv mass fraction of volatiles in gas

yw mass fraction of water in gas

Z stoichiometric component

Greek letters

porosity of clinker bed in cooler

c emissivity of coal particle

cy emissivity of cyclone outer wall

L emissivity of solid particlem,p mass efficiency of the cyclone

residence time of coal particle in calciner, s

g density of air/gas, kg/m3

s density of solids, kg/m3

StephanBoltzmann constant(W/m2 K4)

residence time of raw meal particle in calciner, s

Chemical species

C2S (2CaOSiO2)

C3S (3CaOSiO2)C3A (3CaOAl2O3)

C4AF (4CaOAl2O3Fe2O3)

Acknowledgments

The authors wish to acknowledge financial support provided

by CSIR (under the NMITLI scheme) for this study. The au-

thors would also like to acknowledge many helpful discussions

with Professor Anurag Mehra during the course of this work.

One of the authors, K.S.M is grateful to Council of Scientific

and Industrial Research (CSIR), India for providing financial

support.


18/18


References

Engin, T., Ari, V., 2005. Energy auditing and recovery for dry type cement

rotary kiln systemsa case study. Energy Conversion and Management

46 (4), 551562.

Gupta, A.V.S.S.K.S., Nag, P.K., 2000. Prediction of heat transfer coefficient in

the cyclone separator of a CFB. International Journal of Energy Research

24, 10641079.Hamor, R.J., Smith, I.W., Tyler, R.J., 1973. Kinetics of combustion of

pulverised brown coal char between 630 and 2200 K. Combustion and

Flame 21, 153.

Hindmarsh, A.C., 1983. ODEPACK, a systematized collection of ODE

solvers. In: Stepleman, R.S., et al. (Eds.), Scientific Computing. IMACS

Transactions on Scientific Computation, vol. 1. North-Holland,Amsterdam,

pp. 5564.

Hu, N., Scaroni, A.W., 1996. Calcination of pulverised limestone particles

under furnace injection conditions. Fuel 75, 177186.

Kang, S.K., Kwon, T.W., Kim, S.D., 1989. Hydrodynamic characteristics of

cyclone reactors. Powder Technology 58, 211220.

Kramers, H., Croockewit, P., 1952. The passage of granular solids through

inclined rotarykilns. Chemical Engineering Science 1, 259265.

Lede, J., Li, H.Z., Soulalignac, F., Villermaux, J., 1987. Measurment of solid

particle residence time in cyclone reactor: a comparison of four methods.Chemical Engineering Process 22, 215222.

Li, J.Q., Wei, F., Jin, Y., 2003. Numerical simulation of pulverized

coal combustion and NO formation. Chemical Engineering Science 58,

51615171.

Locher, G., 2002. Mathematical models for the cement clinker burning

process, part 4: grate cooler. ZKG International 55 (6), 4657.

Lu, J., Huang, L., Hu, Z., Wang, S., 2004. Simulation of gassolid, two-phase

flow, coal combustion and raw meal calcination in a pre-calciner. ZKG

International 57 (2), 5563.

Mastorakos, E., Massias, A., Tsakiroglou, C.D., Goussis, D.A., Burganos,

V.N., 1999. CFD predictions for cement kiln including flame modeling, heat

transfer and clinker chemistry. Applied Mathematical Modeling 23, 5576.

Mujumdar, K.S., Ranade, V.V., 2003. Coupling of CFD models with different

time scalesa case study of rotary cement kiln. In FLUENT UsersGroup Meeting held at Pune.

Mujumdar, K.S., Ranade, V.V., 2006. Simulation of rotary cement kilns using

a one-dimensional model. Transactions of IChemE, Part A, Chemical

Engineering Research and Design 84 (A3), 165177.

Mujumdar, K.S., Arora, A., Ranade, V.V., 2006. Modeling of rotary cement

kilns: applications to reduction in energy consumption. Industrial and

Engineering Chemistry Research 45 (7), 23152330.

Nsofor, E.C., Adebiyi, G.A., 2001. Measurements of the gasparticle

convective heat transfer coefficient in a packed bed for high-temperature

energy storage. Experimental Thermal and Fluid Sciences 24 (12), 19.

Peray, K.E., 1984. Rotary Cement Kilns. second ed. Chemical Publishing

Company Ltd., New York.

Spang, H.A., 1972. A dynamic model of a cement kiln. Automatica 8,

309323.

Stanmore, B.R., Gilot, P., 2005. ReviewCalcination and carbonation oflimestone during thermal cycling for CO2 sequestration. Fuel Processing

Technology 86, 17071743.

Warudkar, S., Kumar, G., Mujumdar, K.S., Ranade, V.V., 2005. Modeling of

cement plant calciners. In: CHEMCON 2005, New Delhi.

Watkinson, A.P., Brimacombe, J.K., 1982. Limestone calcination in a rotary

kiln. Metallurgical Transactions B 13B, 369378.

Modeling of clinker cooler

Documents

Transcript of Modeling of clinker cooler