Heat andMass Transfer in Reduction Zone of Sponge Iron Reactor.pdf

7/29/2019 Heat andMass Transfer in Reduction Zone of Sponge Iron Reactor.pdf

http://slidepdf.com/reader/full/heat-andmass-transfer-in-reduction-zone-of-sponge-iron-reactorpdf 1/12

International Scholarly Research Network ISRN Mechanical EngineeringVolume 2011, Article ID 324659, 12 pagesdoi:10.5402/2011/324659

Research ArticleHeat andMass Transfer inReductionZone of Sponge IronReactor

Bayu Alamsari,1 Shuichi Torii,1 Azis Trianto,2 andYazid Bindar2

1 Department of Mechanical System Engineering, Kumamoto University, Kumamoto 860-8555, Japan 2 Department of Chemical Engineering, Bandung Institute of Technology, Bandung 40132, Indonesia

Correspondence should be addressed to Azis Trianto, [email protected]

Received 13 March 2011; Accepted 14 May 2011

Academic Editor: S.-H. Chuang

Copyright © 2011 Bayu Alamsari et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Numerical prediction is performed on reduction zone of iron ore reactor which is a part of counter current gas-solid reactor forproducing sponge iron. The aim of the present study is to investigate the eff ect of reduction gas composition and temperature onquality and capacity of sponge iron products through mathematical modeling arrangement and simulation. Simultaneous massand energy balances along the reactor lead to a set of ordinary diff erential equation which includes kinetic equations. Kineticequations of reduction of hematite to iron metal, methane reforming, and water gas shift reaction are taken into account in themodel. Hydrogen and carbon monoxide are used as reduction gas. The equations were solved by finite element method. Predictionshows an increase in H2 composition while an attenuation of CO produces higher metallization degree. Metallization degree isalso increased with an increase in gas inlet temperature. It is found that reduction gas temperature over 973◦C (1246 K) is notrecommended because the formation of sticky iron will be initiated.

1. Introduction

Sponge iron product quality in direct Reduced iron (DRI)plant can be increased by rising reduction gas temperatureand composition of reactant in reduction gas. The risingof temperature will accelerate reaction flow rate which inturn can raise reaction conversion as shown by Zhang andOstrovski [1]. However, the temperature rising can also bringdisadvantage due to the formation of total carbon on theproduct. The increasing of reactant composition which isconsisted of H2 and CO in reduction gas will also increaseproduct quality as reported by many researchers (e.g., [2–

4]). The optimum composition of reduction gas must bedetermined to obtain the highest quality.

The investigation of temperature and composition of reduction gas is carried out by analyzing the performance of sponge iron reactor which in turn can predict the profitableoptimum condition of reactor. However, since the reactoranalysis is difficult to be executed directly in the field plant, itwill need a simulator as support instrument. This simulatorcan be designed by arrangement of heat and mass transferequation inside the reactor to reproduce reactor data fromthe field. By solving heat and mass transfer equations, theeff ect of reduction gas on the performance of reactor isstudied.

Some iron ore reactor models are proposed in a few liter-atures. Iron ore reactor model proposed by Aguilar et al. [5]is referred for fixed bed reactor. The model is derived basedon unsteady state condition. Parisi and Laborde [6] proposeda model for moving bed counter current reactor. The modelis limited on reduction zone of iron ore reactor, and only reduction reactions are considered in the model, in whichplug flow is used as an approach for the model.

Another related model is formulated on direct reductionshaft furnace for producing sponge iron from iron ore as pro-posed by Takenaka et al. [7]. In their model, reduction rate

equations are derived from the three-interface model whichinvolvesboth mass and heat balances. Srinivasan [8] developsa model for reduction iron oxides by carbon. This model inwhich only the global direct reduction reactions are takeninto account and methane reforming and water gas shiftreaction are ignored, is formulated on a circulating fluidizedbed reactor.

The kinetics of direct reduction reaction plays an impor-tant role in sponge iron reduction zone mathematical model.Unreacted core model proposed by Levenspiel [9] is widely used for kinetics model of iron reduction in which the reac-tion occurs first at the outer skin of the particle. The zoneof reaction then moves into the solid and may leave behind



2 ISRN Mechanical Engineering

completely converted material and inert solid. Thus, at any time, there exists an unreacted core of material which shrinksin size during reaction.

Mondal et al. [10] propose the another kinetics model foriron reduction. Here, the kinetic constants are derived basedon Arrhenius equation and Cementite formation reaction

was also taken into account in their model. Note that thismodel is only applicable for iron reduction using CO asreducer. Iguchi and Yokomoto [11] also suggest kineticsmodel based on Arrhenius equation but cementite formationwas ignored.

In the present study, heat and mass transfer in spongeiron reactor reduction zone are studied. Here, the model canbe used to explore the performance of sponge iron reactor.In particular, the eff ect of reduction gas composition andtemperature on quality and capacity of sponge iron productsis precisely disclosed, because kinetics equation of reductionof hematite to iron metal, methane reforming, and water gasshift reaction are taken into account in the model.

2. ReactorModeling

Iron ore reactor is a moving bed reactor to produce spongeiron product from iron ore. Gas and solid phases are passedin counter current. In general, sponge ore reactor is dividedby 3 zones, which are referred to as reduction, isobaric,and cooling zones. The simple scheme of iron ore reactoris shown in Figure 1. In the reduction zone, there are somereactions occurring between reduction gas and iron orepellets. These reactions which are called reduction reactionwill extract iron metal from the ore to create sponge ironproduct. These reactions occur on the surface of iron orepellet. In general, hydrogen and carbon monoxide are used

as reducer and the corresponding reduction reactions areshown as follows:

3Fe2O3 + CO(H2) ⇐⇒ 2Fe3O4 + CO2(H2O) (1)

Fe3O4 + CO(H2) ⇐⇒ 3FeO + CO2(H2O) (2)

FeO + CO(H2) ⇐⇒ Fe + CO2(H2O). (3)

Methane reforming and water gas shift reactions alsooccur in the gas phase based on the composition of reductiongas and temperature through reaction as follows:

CH4 + H2O⇐⇒

CO2 + H2 (4)

CO2 + H2 ⇐⇒ CO + H2O. (5)

Mathematical model of reduction zone in iron ore reac-tor is arranged on the gas and solid phases. Heat and masstransfer equations are formulated to perform the tempera-ture and concentration of gas and solid phases, respectively.The modeling is also carried out by employing kineticsmodel. Note that iron reduction, methane reforming, andwater gas shift reactions are taken into account in the model.

The mass and energy equations are formulated by assuming (i) steady-state operating conditions, (ii) no heatloss over the wall of reactor reduction zone, (iii) the

iron ore pellet consumption is governed by the unreactedshrinking core model, (iv) the enthalpy is calculated basedon temperature changes, and (v) the plug flow is used toapproach a model for gas and solid phases. By consideringthe above assumptions, heat and mass balance equations canbe stated as follows:

ugasdC CH4

dz = r methanation

CH4, H2, T gas

, (6)

ugasdC H2

dz = r WGSR

CO, T gas

+ r methanation

CH4, H2, T gas

− n p · r H2-reduction(H2, T solid),

(7)

ugasdC CO

dz = r WGSR

CO, T gas

+ r methanation

CH4, H2, T gas

− n p · r CO-reduction(CO, T solid),

(8)

usoliddC Fe2O3

dz = n p · r Fe2O3 (CO,H2, T solid), (9)

usoliddC Fe3O4

dz = n p · r Fe3O4 (CO,H2, T solid), (10)

usoliddC FeO

dz = n p · r FeO(CO,H2, T solid), (11)

usoliddC Fe

dz = n p · r Fe(CO,H2, T solid), (12)

dT gas

dz =

Asp · hT solid − T gas

M gas · Cpgas

+R f gas

[H ir i]

M gas ·Cpgas,

(13)

dT solid

dz =

Asp · h ·T solid − T gas

M solid · Cpsolid

−

[H ir i]

M solid · Cpsolid.

(14)

Equations (6)–(8) are mass balance equations for gas phase,while (9)–(12) are for solid phase. Heat balance equations forgas and solid phase are expressed by (13) and (14), respec-tively. Notation

Aspis pellets surface area per unit reactor

volume, h is convection heat transfer coefficient, and M ismolar flow rate.

Heat capacities (Cp) of gas and solid phases in (13) and(14) are calculated based on the temperature changes in eachphase. On the gas phase, heat capacity value is calculatedbased on the mixing of heat capacity of CO and H2. Theothercomponents such as H2O, CH4, CO2, and N2 are omittedbecause they have small fraction compared to CO and H2.The heat capacity of gas phase is evaluated as follows:

Cpgas =

C CO

C CO + C H2

CpCO

+

C H2

C CO + C H2

CpH2

, (15)



ISRN Mechanical Engineering 3

Reductionzone

Isobariczone

Coolingzone

Coolinggas input

Cooling

gas output

Sponge ironproduct

Sponge ironfrom isobaric zone

1

23

7 8

9

Reductiongas waste Iron ore

Reduced iron

Makeup

cooling gas

45

6

10

Reduction gas

Figure 1: Scheme of sponge iron reactor.

where

Cp(CO or H2) = a1T gas − a2T 2gas + a3T 3gas − a4T 4gas + a5T 5gas.

(16)

Here, constant values of a1, a2, a3, a4, and a5 for CO andH2 are defined by Reklaitis [12]. Heat capacity of solid phaseis calculated based on heat capacity of Fe. Heat capacity parameter of Fe as obtained from Green and Perry [13] isshown as follows:

Cp(Fe) = 0.0029T solid − 5E−6T 2solid + 3E−9T 3solid. (17)

Standard reaction enthalpy (H i) and kinetic equation (r i)in (13) and (14) are calculated based on reactions in eachphase. Sigma notation (

H ir i) in (13) is calculated through

methanation and water gas shift reactions, while (

H ir i)in (14) is calculated by reduction reactions of CO and H2.The formulation of sigma notation for (13) is described asfollows:

H ir i

=

H R(298)-methanation +

T o

T gas

CpRdT

r methanation

+

H R(298)-WGSR + T o

T gas CpRdT

r WGSR

,

(18)

while formulation of sigma notation for (14) is as follows:H ir i

=

H R(298) CO-reduction +

T o

T solid

CpRdT

r CO-reduction

+

H R(298) H2-reduction +

T o

T solid

CpRdT

r H2-reduction

.

(19)

The values of standard reaction enthalpy at 298 K (H R(298))for each reaction are obtained from Chemical Sciences [12].Reaction heat capacity (CpR) is also calculated for eachreaction by using

CpR =i

biCpi-reactant − j

b jCp j-product, (20)

where i and j refer to reactant and product number, respec-tively. Notation b is stoichiometric coefficient for each reac-tant and product.

Heat transfer coefficient, h, is determined by using thecorrelation proposed by Furnas [14]. He investigates heattransfer value from a stream of air to a bed of iron pelletscovered with a thin coating of iron oxides. The temperatureexperiments are set up to 1023 K. The equation proposed by Furnas [14] is as follows:

h = 6.91× 10−3 ×G0.75 f × T ×G−1.56 × 5.67826. (21)

Notation r in (6)–(14) is referred to as kinetic expressionof every reaction which occurs in reduction zone. There arethree kinetic equations involved in the model, namely, akinetic equation of methanation reaction, a kinetic equationof water gas shift reaction, and a kinetic equation of reduc-tion reaction. Kinetic expression of methanation reaction(r methanation) is adopted from Munster and Grabke [15] givenby

r methanation = f methanationkoa0 exp

−E+

A

RT

pCH4√ pH2

. (22)

Here, the values of activation energy (E A) and preexponentialconstant (ko) are the same as those adopted in the literature

[15]. Constant, a0, is expressed as pH2O/ pH2. Note that cor-rection factor f methanation proposed here is added so as tocorrect the value of preexponential constant.

Many researchers propose kinetic equations for water gasshift reaction. The validity of kinetic model is to get resultswhich have a good agreement with reference data. Somekinetics models are valid for low-temperature reaction whilesome others are not. Bustamante et al. [16] carried out onthe temperature range of 1148–1198 K whose condition issimilar to that of the present study. Their kinetic model isshown as follows:

r WGSR = f WGSR · ko · exp

−E+

A

RT

C H2

0.33

C CO2. (23)

Here, the value of ko is corrected by correction factor f .Kinetic equations of reduction reaction in iron ore are

created based on shrinking core model for spherical particles.The formulations of kinetics equation are derived basedon the model proposed by Levenspiel [9]. Although many literatures propose the models such as Mondal et al. [10],Iguchi and Yokomoto [11], and Shi et al. [17]; amongthem, Levenspiel formulates the model in more detail by considering every step which is close with reality. Figure 2shows the visualization of shrinking core model for ironreduction. Basically, 5 steps take place during reaction, asexplained in Levenspiel [9].




Gas film

Ash

0 R pR p r cr c

Fe2O3

Fe3O4

FeOFe

CO and H2

Reactant

CO2 and H2O

Product

Enter the pellet by

diff using through

gas film and ash

Leave the pellet by

diff using through

gas film and ash

Figure 2: Shrinking core model.

Gas film

Δx

C CO or H2

C ash

Qgas Qash

Figure

3: Diff

usion through gas film.

(a) Diff usion of gaseous reactants (CO and H2) throughthe film surrounding the particle to the surface of thesolid.

(b) Diff usion of CO and H2 through the blanket of ash tothe surface of the unreacted core of iron pellet.

(c) Reaction of CO and H2 with iron ore at reactionsurface through

3Fe2O3 + CO(H2) −→ 2Fe3O4 + CO2(H2O) (24)

Fe3O4 + CO(H2) −→ 3FeO + CO2(H2O) (25)

FeO + CO(H2) −→ Fe + CO2(H2O). (26)

(d) Diff usion of gaseous products (CO2 and H2O)through the ash back to gas film.

(e) Diff usion of gaseous products through the gas filmback into the main body of fluid.

The kinetic equation of each step must be combined toget overall kinetic expression for each component involvedin reduction reaction. Therefore, the expression of kineticequation of each step must be defined in the same manner.

Since the rate of mass transfer (gas diff usion) on steps (a),

(b), (d), and (e) can be defined as the flow of material unitsurface, or

Qdiff usion = −1

S

dN gas

dt , (27)

the reaction step (step (c)) must similarly be defined as

Qreaction = −1

S

dN idt

. (28)

Based on those steps, we can formulate the overall kineticequation based on 3 main equations; those are as follows.

(1) Di ff usion through Gas Film Control (Steps (a) and (e)). If gas CO or H2 diff uses through a stagnant film onto surface of ash as shown in Figure 3, the flux of material can be definedas the flow of gas unit surface, or

Qgas = −1

S

dN CO or H2

dt = D

ΔC

Δr = D

Δr

C CO or H2 −C ash

= kfilm

C CO or H2 −C ash

,

(29)

where S is surface area (m2), N is mole amount of COor H2 in mole (kfilm) is gas film kinetic constant (m/sec),C is concentration (mole/m3), and D is molecular diff usion

coeffi

cient (m2

/sec). The same equation can also be appliedfor diff usion of gas product (CO2 and H2O).On the ash surface, the flux diff usion can also be stated as

Qash = −1

S

dN CO or H2

dt = kashC ash, (30)

where kash is kinetic constant on ash. Since C ash is difficultto be measured, this variable must be disappeared in (29) or(30). At steady state, the flow rate to the surface is equal tothe reaction rate at the surface, thus

Qgas = Qash,

kfilmC CO or H2

−C ash =

kashC ash.(31)




Therefore,

C ash =kfilm

kfilm + kashC CO or H2 . (32)

If we insert (32) to (29) or (30), we will get

Qgas = −1

S

dN CO or H2

dt = kashk

filmkfilm + kash

C CO or H2

= kGASC CO or H2 ,

(33)

where kGAS is overall kinetic constant on gas film. Equation(33) then will be used as kinetic expression of diff usionthrough gas film control.

(2) Di ff usion through Ash Layer Control (Steps (b) and (d)).If the flux of gas within the ash layer is expressed by Ficks law for equimolar counter diff usion, then we will have

Q= −

1

S

dN CO or H2

dt =℘e

dC CO or H2

dr

, (34)

where ℘e is eff ective diff usion coefficient of gaseous reactantin the ash layer.

If we define S = 4πr 2, we will get

−dN CO or H2

dt = 4πr 2 ·℘e

dC CO or H2

dr ,

−dN CO or H2

dt

Rc

R p

dr

r 2= 4π ·℘e

0C CO or H2 at R p

dC CO or H2 .

(35)

Then, by integrating (34), we get

−dN CO or H2

dt 1

Rc −1

R p = 4π ℘eC CO or H2 . (36)

If we multiply both sides with 1/ A p where A p is pellet surfacearea and define A p = 4πR2

p where R p is pellet radius, we get

− 1

A p

dN CO or H2

dt

1

RC − 1

R p

= 1

4πR2 p

4π ℘eC CO or H2= 1

R2 p℘eC CO or H2

(37)

or

− 1

A p

dN CO or H2

dt

= 1

R2 p

(1 /Rc)−

1 /R p

×℘eC CO or H2 .(38)

Equation (38) then will be used as kinetic expression of diff usion through ash layer control.

(3) Chemical Reaction Controls (Step (c)). There are 6 maincomponents which involve in reduction reaction; those arehematite (Fe2O3), magnetite (Fe3O4), wustite (FeO), metaliron (Fe), carbon monoxide (CO), and hydrogen (H2). Thereactions are assumed pseudofirst order; that is, the rate of reaction for each component is formulated as follows.

(a) Fe2O3:

−1

S

dN Fe2O3

dt = − 1

4πR2c

dN Fe2O3

dt = k1C Fe2O3 , (39)

where Rc is unreacted core radius in pellet.If we multiple each side by 1 /A p and define A p

=4πR2

p,

we get

− 1

A p

dN Fe2O3

dt = R2

c

R2 p

k1C Fe2O3 . (40)

(b) Fe3O4:

− 1

A p

dN Fe3O4

dt = R2

c

R2 p

− X · k1C Fe2O3 + k2C Fe3O4

. (41)

(c) FeO:

−1

A p

dN FeO

dt =R2

c

R2 p−Yk2C Fe3O4 + k3C FeO

. (42)

(d) Fe:

− 1

A p

dN Fe

dt = R2

c

R2 p

(−Zk3C FeO). (43)

(e) CO:

− b

A p

dN CO

dt = R2

c

R2 p

V 1k1C Fe2O3 + k2C Fe3O4 + k3C FeO

. (44)

(e) H2:

− b

A p

dN H2

dt = R2

c

R2 p

V 2k1C Fe2O3 + k2C Fe3O4 + k3C FeO

. (45)

Notations X , Y , and Z shown in (41), (42), and (43), respec-tively, represent the weight coefficient, while V 1 and V 2 on(44) and (45) represent equilibrium conditions.

Kinetic reaction constants k1, k2, and k3, respectively,refer to reaction as

3Fe2O3 + CO(H2)k1−→ 2Fe3O4 + CO2(H2O) (46)

Fe3O4 + CO(H2)k 2

−→3FeO + CO2(H2O) (47)

FeO + CO(H2)k3−→ Fe + CO2(H2O). (48)

If we combined the 3 main equations shown above, we get theindividual equation for each component involved in reduc-tion reaction as written below.

(a) Fe2O3:

− 1

A p

dN Fe2O3

dt = bkGASC CO & H2 +

b℘eC CO & H2

R2 p

1 /Rc − 1 /R p

+ bR2

c

R2 p

k1(CO & H2)C Fe2O3

(49)




or

−r Fe2O3 = A pbkGASC CO & H2 + A pb℘eC CO & H2

R2 p

1 /RC − 1 /R p

+ A pbR2

C

R2 p k1(CO & H2)C Fe2O3.

(50)

(b) Fe3O4:

− r Fe3O4

= A pbkGASC CO & H2 + A pb℘eC CO & H2

R2 p

1 /RC − 1 /R p

+ A pbR2

C

R2 p

− X · k1(CO & H2)C Fe2O3 + k2(CO & H2)C Fe3O4

.

(51)

(c) FeO:

−r FeO = A pbkGASC CO & H2 + A pb℘eC CO & H2

R2 p

1 /RC − 1 /R p

+ A pbRC

R2 p

−Y · k2(CO & H2)C Fe3O4

+k3(CO & H2)C FeO

.

(52)

(d) Fe:

−r Fe

= A pbkGASC CO & H2 +

A pb℘eC CO & H2

R2 p

1 /RC − 1 /R p

+ A pbRC

R2 p

−Z · k3(CO & H2)C FeO

.

(53)

(e) CO:

− r CO−reduction

= A pbkGASC CO + A pb℘eC CO

R2 p

1 /RC − 1 /R p

+ A pbR2

C

R2 p

V 1k1(CO)C Fe2O3 + k2(CO)C Fe3O4 + k3(CO )C FeO.(54)

(f) H2:

− r H2−reduction

= A pbkGASC H2 + A pb℘eC H2

R2 p

1 /RC − 1 /R p

+ A pbR2

C

R2 p

V 2k1(H2)C Fe2O3 + k2(H2)C Fe3O4 + k3(H2)C FeO

.

(55)

Here, kinetic equations r i shown in (50)–(55) are formulatedbased on reaction rate on 1 single pellet. Therefore, they mustbe multiplied by the number of pellets per unit volume (n p)to get the overall reaction rate.

Reaction constants (k1, k2, and k3) in (50)–(55) are mod-eled based on Arrhenius method which is stated as

ki = ko · exp

−E+

A

RT

. (56)

Here, as for two variables unknown in (56), those are pre-exponential constant ko and activation energy E A. Note thatthe corresponding values are diff erent for each reductionreaction by CO and H2. Many literatures focus on thecomparison of E A values rather than ko, because the variationof ko values is large and highly dependent on researchcondition. The value of activation energy (E A) is importantto get good result. The ko and E A values used in this researchare adopted from Aguilar et al. [5]. The formulations of eachki are shown in Table 1.

Overall kinetic constant on gas film kGAS is calculatedbased on formulation proposed by Parisi and Laborde [6] asstated as

kGAS = 0.00225 exp

−14700

82.06T

, (57)

while eff ective diff usion coefficient ℘e is calculated using

℘e =ε p ·D

τ p. (58)

Considering the unreacted core model, it is possible torelate the radius of the unreacted core (r c) with the solid

conversion (Co) by

RC =

R3 p −

Co · M wn p4πρ

1 / 3

, (59)

where

Co =C 0CO −C CO

+C 0H2

− C H2

. (60)

Simulation is carried out on sponge iron production capacity of 2500 tone/day. Reduction gas flow rate required is about125000 m3/hour, while leaking cooling gas flow rate fromisobaric zone is 4700 m3/hour.

Calculation of the model is executed by using finiteelement method. The calculation is divided by many seg-ments along height of reduction zone. Figure 4 shows thesimulation algorithm to solve mathematical equation on thecooling zone. Flow numbers shown in this figure refer toFigure 1. By using data at flow 1 and 6, we can then solve themathematical model simultaneously. To check the validity of the model, we have to compare the result at flow 4 and 5 withdata from the plant. If their temperature and concentrationof each flow are not in agreement with reference, we haveto adjust correction factor R f for temperature and f k forconcentration. There are 1 R f and 5 f k involved in thecalculation at reduction zone. Factor R f is diff erentiated




Table 1: Reaction constants formulation for kinetic equation of iron reduction.

CO H2

Fe2O3 → Fe3O4 ki(CO) = 9 · exp[−49884 /RT ] ki(H2 ) = 45 · exp[−49884 /RT ]

Fe3O4 → FeO ki(CO) = 0.072 · exp[−21616 /RT ] ki(H2 ) = 0.36 · exp[−21616 /RT ]

FeO → Fe ki(CO) = 0.036 · exp[−21616 /RT ] ki(H2) = 9 · exp[−21616 /RT ]

Table 2: Correction factor value used in mathematical model of reduction zone.

No. NoteMethane

reformingWater gas shift

reaction

(1) Correction factor f — 0.036

(2) Reaction constant rate ∗0.00199 —

(3) R f 0.02

(4)

X 0.02

Y 40

Z 14800

V 1 4.6E−6

V 2 5.1E−6∗

Reaction constant rate (mol/m2·s·Pa0.94).

based on the temperature phase, while f k is diff erentiatedbased on kinetic equation of each reaction.

Reaction constant rate value for methane reforming isnot explained in literature. The value shown in Table 2 isobtained from the calculation after fitting with the referencedata.

3. Results andDiscussion3.1. Temperature, Gas, and Solid Profile along Reduction Zone.Table 3 resumes the comparison between reference data andsimulation. Reference data are taken from Project FinalReport [18]. From this table, one observes that simulationprecisely predicts reference data but not calculation result onthe other zone. Here, root mean square error (RMSE) showsa value around 0.19. Gas composition data for flow number 1are not available in the plant, but they can be calculated frommass balance resulted from mixing between reduction gasinlet (flow number 2) and leaking cooling gas (flow number3). Leaking cooling gas is obtained from Alamsari et al. [19].

Reduced iron data (flow number 6) at the outlet of

reduction zone are also unavailable in the plant. However,the simulation results from the calculation of reduction zonealready are in good agreement with simulation result pro-duced from the calculation in isobaric zone [19]. Alamsariet al. propose model for isobaric and cooling zone of thesame reactor. They found that metallization degree of theproduct reduces around 1.72% after leaving isobaric andcooling zone. As metallization degree of sponge iron productfrom the plant data is stated about 92.49%, it implies that theproduct (reduced iron) before entering isobaric and coolingzone has metallization degree of 94.21%. This value agreeswith simulation result produced from this research. Reducediron shown in Table 3 produces metallization degree around

Table 3: Comparison between reference data and simulation.

Flow Note Plant dataSimulation results

Reductionzone

Isobariczone

(1)

Reduction gas in the bot-tom of reduction zone

Temperature (K) 1193 1193

CH4 (% vol.) 2.9

H2 (% vol.) 42.1

CO (% vol.) 9.58

CO2 (% vol.) 2.26

N2 (% vol.) 3.72

H2O (% vol.) 1.36

(2)

Reduction gas inlet

Temperature (K) 1203

CH4 (% vol.) 3.7

H2 (% vol.) 71.86

CO (% vol.) 16.5

CO2 (% vol.) 3.8

N2 (% vol.) 1.8

H2O (% vol.) 2.34

(3)

Leaking gas from isobariczone

Temperature (K) 1139

CH4 (% vol.) 50.13

H2 (% vol.) 45.63

CO (% vol.) 0.983

CO2 (% vol.) 3.03

N2 (% vol.) 1.14

H2O (% vol.) 0.12

(4)

Reduction gas wastes

Temperature (K) 657.5 657.5

CO2 (% vol.) 9.0 8.5

CH4 (% vol.) 4.4 4.5

CO (% vol.) 12.1 11.7

H2O (% vol.) 23.4 23.9

N2 (% vol.) 12 1.7

H2 (% vol.) 49.7 49.5

94.22%. Alamsari et al. [19] also show solid temperatureprofile along isobaric and cooling zone. Reduced irontemperature was obtained about 1169 K. This value is closewith simulation result shown in Table 3.




Reduction

Input

T 1

M 1H2

M 1CO

M 1CO2

M 1H2O

M 1CH4

M 1N2

u1

T 6

M 6Fe2O3

M 6Fe3O4

M 6FeO

M 6Fe

M 6gangue

M 6Fe2+

u6

Solve momentum, ma ss , andenergy balance equations

simultaneously starting fromthe bottom of reduction zone

Results

T 4

M 4H2

M 4CO

M 4CO2

M 4H2O

M 4CH4

M 4N2

u4

T 5

M 5Fe2O3

M 5Fe3O4

M 5FeO

M 5Fe

M 5gangue

M 5Fe2+

u5

Adjust R f

Adjust f k

Finish

Yes

No

Yes

NoAre T 4 ≈ T 4ref and T 5 ≈ T 5ref ?

Are M 4i ≈ M 4i-ref and M 5i ≈ M 5i-ref ?

Figure 4: Simulation algorithm on the reduction zone.

0 1 2 3 4 5 6

Depth of reduction zone (m)

100

200

300

400

500

600

700

800

900

1000

1200

1100

T e m p e r a t u r e ( K )

SolidGas

0

(a)

CO2

H2O

CH4CO

H2

80

70

60

50

40

30

20

10 G a s c o n c e n t r a t i o n ( % v o l u m e )

0

N2

0 1 2 3 4 5 6


(b)

S o l i d c o n c e n t r a t i o n ( % m a s s )

0

20

40

60

80

100

Fe2O3

Fe3O4

FeO

Fe

Gangue

Fe2+

0 1 2 3 4 5 6

Solid flow Gas flow


(c)

Figure 5: Profiles of many variables as a function of the depth of reduction zone; (a) gas and solid temperature; (b) gas concentration; (c)solid concentration.




Are M 5i ≈ M 5i-ref ?

ref ?Is T 5 ≈ T 5No

NoYes

YesFinish

T 6

M 6iu6

T 1

M 1i

u1

Flow 4

Flow 5T 5

M 5iu5

T 4

M 4i

u4

Adjust M 6iAdjust T 6

Input

Solvemathematical

model

simultaneously

Results

Reductionzone

Reduction gas input

Reduction

gas wasteIron ore

Reduced iron

Cooling gas leak

New

1

3

6

54

Figure 6: Simulation algorithm of reduction gas temperature eff ect on metallization degree of reduced iron.

Reduction gas inlet temperature (K)1160 1180 1200 1220 1240 1260 1280

0.92

0.94

0.93

0.97

0.96

0.95

M e t a l l i z a t i o n d e g r e e o f r e d u c e d i r o n

Figure 7: Metallization degree as function of gas inlet temperature.

The composition data of iron oxides are only providedfor hematite (Fe2O3) as the main component, while theother components such as magnetite (Fe3O4) and wustite(FeO) are ignored because they have very small compositioncompared with hematite. However, since the main goal inthis research is to model the whole iron reactor and kineticsequation is only small part of the model, the concentrationof magnetite (Fe3O4) and wustite (FeO) will not aff ect the

result.Figure 5(a) illustrates temperature profiles along reduc-

tion zone for gas and solid phases. Inlet and outlet temper-atures are in good agreement with the reference data. Oneobserves that temperature profiles along reduction zone areincreased with an increase in gas and solid temperature fromthe direction of solid flow. The corresponding concentrationprofiles for gas phase are shown in Figure 5(b). Except nitro-gen, other components show composition changes whichare caused by reactions in the gas phase. Gas mole balancereactions for methane reforming, iron reduction, and watergas shift reaction already have an agreement. Figure 5(c)depicts solid concentration profile along reduction zone. It is

observed that, at the beginning of reduction zone inlet, totalFe in the ore is dominated by Fe2O3, while as reactions occur,Fe2O3 is slowly reduced by reduction gas to release oxygen.

3.2. Analysis the E ff ect of Temperature and Composition of Reduction Gas on Product Quality. Reduction gas as reactantin reduction reaction plays an important role in an enhance-ment of product quality. The optimum condition of reduc-tion gas must be investigated to get the high quality. Theeff ect of reduction gas on product quality is investigated by studying the eff ects of temperature and composition of reduction gas on metallization degree. The investigation iscarried out by numerical simulation using the mathematicalmodel arranged before.

3.3. Temperature. An eff ect of temperature on metallizationdegree is numerically investigated by varying inlet tempera-ture of reduction gas on the range of 1170–1260 K with anincrement around 10 K. The calculation results on reductionzone (Table 3) are used as the basis values of this simulation.The inlet temperature in this simulation is referred to as thetemperature mixing of leaking cooling gas at flow 3 andreduction gas at flow 2 (see Figure 1).

Figure 6 shows the algorithm of reduction gas tempera-

ture eff ect on metallization degree. The calculation is startedfrom the bottom of reduction zone. Data at flow 1 and 6 areused as input in the calculation. The simulation is carriedout for each value of gas inlet temperature, and new dataat flow 4 and 5 of each value are obtained. The result atflow 5 must be in good agreement with the plant data sincethis flow is the real input in iron reactor which is replacedby flow 6. Therefore, we have to adjust temperature (T )and concentration ( M ) at flow 6 until temperature (T ) andconcentration ( M ) at flow 5 agree with reference data.

Figure 7 illustrates the relation between metallizationdegree of reduced iron with reduction gas inlet temperature.It is observed that reduction reaction of hematite produces




ref ?Is T 5 ≈ T 5No

No

Yes

Finish

Flow 4

Flow 5

Adjust M 6i

T 6

M 6i

u6

T 1

M 1H2

M 1

CO

M 1H2O

M 1CO2

M 1CH4

M 1N2

u1

New

New

Input

Solve

mathematical

model

simultaneously

Results

Reduction

zone

Reduction gas input

Reduction

gas wasteIron ore

Reduced iron

Cooling gas leak

T 5

M 5iu5

T 4

M 4i

u4

Are M 5i ≈ M 5i-ref ? Yes

Adjust T 6

3

4

1

6

5

Figure 8: Simulation algorithm of reduction gas composition eff ect on metallization degree of reduced iron.

M e t a l l i z a t i o n d e g r e e

1

0.9

0.75

0.8

0.85

0.95

Ratio H2/CO (%)

52/35 56/31 60/27 64/23 68/19 72/15 76/11

Figure 9: Metallization degree as function of ratio H2–CO.

higher metallization degree as an increase in reduction gasinlet temperature. This occurs because the increase in reduc-

tion gas inlet temperature causes both an enhancement of reaction temperature and an accelerate reaction flow rate.Zhang and Ostrovski [1] obtain the same result when they investigated iron ore reduction by H2–CH4–Ar gas mixtures.They disclose that the rate of iron ore reduction is intensifiedwith an increase in temperature.

Although an increase in temperature can be an alterna-tive way to get a better metallization degree, this increasing islimited by physical properties of iron. Diaz et al. [20] suggesta setting gas inlet temperature high enough to allow forliquid-phase migration of the metal iron within the pellets.However, temperature must be also maintained below thepoint at which the pellets became sticky. Sticky irons bring

troubles because they will stimulate an agglomeration,sintered to each other and form accretions on the fur-nace/reactor walls. Based on the plant data, ring reduction

temperature over 1153 K will initiate formation of sticky iron. It implies that reduction gas inlet temperature shouldbe below 1246 K to avoid this problem.

3.4. Composition. In general, hydrogen and carbon monox-ide are good reducers for iron ore. The change in their com-position in reduction gas will aff ect quality of the product.Simulation of the eff ect of composition changes on metal-lization degree is performed on various composition of H2

and CO. Composition of other gasis maintained on the sameratio.

Figure 8 shows the simulation algorithm of reduction gascomposition eff ect on metallization degree. Note that correc-

tion factor and other values resulted from the calculation inreduction zone are used as the basis of calculation. By usingdata at flow 1 and 6 as calculation input, we can solve themodel and get the new value for each data at flow 4 and 5.

Figure 9 depicts metallization degree as a function of ratio H2–CO. One observes that an increase in H2 com-position while an attenuation of CO will produce highermetallization degree. This implies that H2 is better reducerthan CO. The same conclusion is reported by Bonalde et al.[2], El-Geassy and Rajakumar [3], and Pineau et al. [4].Bonalde et al. [2] analyze metallization degree produced by using 100% H2, 100% carbon monoxide, and midrex gas(55% H2-34% CO) as reducer, respectively. They reported




that the fastest reaction occurs when H2 is used and theslowest rate is observed when CO is used, while midrex gas isintermediate. When El-Geassy and Rajakumar [3] modeledthe reduction of iron starting from wustite, they came outwith the same conclusion. Pineau et al. [4] dealt with thereduction of Fe2O3 in the temperature range of 220–680◦C.

They found that the rate of reduction of iron oxide with H2is systematically higher than that obtained by CO.

4. Conclusions

Reduction zone of iron reactor has been simulated. Kineticsequation of reduction of hematite to iron metal, methanereforming, and water gas shift reaction are taken into accountin the model. Simulation results are in good agreement withthe reference data.

Temperature profiles along reduction zone show an in-crease in gas and solid temperature from the direction of solid flow. Gas concentration profiles show compositionchanges which are caused by reactions in the gas phase. Gasmole balance reaction for methane reforming, reduction,and water gas shift reaction already have an agreement.

At the beginning of reduction zone inlet, solid concen-tration profile shows that total Fe in the ore is dominatedby Fe2O3. As reactions occur, Fe2O3 is slowly reduced by reduction gas to release oxygen. Reduced iron from reductionzone produces metallization degree until 94.22%.

When an attenuation of CO yields, an increase in H2

composition produces higher metallization degree. Metal-lization degree is also increased with an increase in gas inlettemperature. Reduction gas temperature over 1246 K is notrecommended because the formation of sticky iron will beinitiated.

Nomenclature

Asp : Pellets surface area per unit reactorvolume, m2/m3

b: Stoichiometr ic coefficientsC : Concentration, mole/m3

Cp: Heat capacity, J/moleKD: Molecular diff usivity of oxygen in gasDn: Eff ective diff usivity of oxygenε p: Pellets porosity E A: Activation energy, J/mole f : Correction factor

M : Molar flow rate, mole/sec·m2

M w: Molecular weight, kg/molenpellet: Number of pellets per unit volume

(1/m3)h: Convection heat transfer coefficient,

J/s·m2·KH R(298 K): Standard reaction enthalpy, J/molek: Preexponential constant mole

C/sec·m3·PaL: Length of reactor cooling zone, m p: Pressure, PaQ: Flux of material, mole/cm2·secr : Reaction rate, mole/m3

·sec

R: Gas constant, Pa·m3/mole·KRc: Radius of unreacted core, mR p: Pellets radius, mR f : Relaxation factorT : Temperature, Kτ p: Pore structure tortuosity

u: Superficial velocity, m/sV : Coefficients depending on weightrelationships and equilibrium conditions

X , Y , Z : Weight coefficientsz : Space variable inside reactor, msolid: SolidR: Reaction.

Superscripts

+: Forward reaction−: Backward reaction.

Subscripts1: Reduction reaction by H2

2: Reduction reaction by CO3: Methane reforming reaction4: Water gas shift reactioni: ith reaction.

References

[1] J. Zhang and O. Ostrovski, “Iron ore reduction/cementation:experimental results and kinetic modelling,” Ironmaking and Steelmaking , vol. 29, no. 1, pp. 15–21, 2002.

[2] A. Bonalde, A. Henriquez, and M. Manrique, “Kinetic analysis

of the iron oxide reduction using hydrogen-carbon monoxidemixtures as reducing agent,” ISIJ International , vol. 45, no. 9,pp. 1255–1260, 2005.

[3] A. A. El-Geassy and V. Rajakumar, “Gaseous Reduction of Wustite with H2, CO and H2-CO mixtures,” Transactions of the Iron and Steel Institute of Japan, vol. 25, pp. 449–458, 1985.

[4] A. Pineau, N. Kanari, and I. Gaballah, “Kinetics of reductionof iron oxides by H2. Part I: low temperature reduction of hematite,” Thermochimica Acta, vol. 447, no. 1, pp. 89–100,2006.

[5] J. Aguilar, R. Fuentes, and R. Viramontes, “Simulation of ironore reduction in a fixed bed,” Modelling and Simulation in

Materials Science and Engineering , vol. 3, no. 2, pp. 131–147,1995.

[6] D. R. Parisi and M. A. Laborde, “Modeling of counter currentmoving bed gas-solid reactor used in direct reduction of ironore,” Chemical Engineering Journal , vol. 104, no. 1–3, pp. 35–43, 2004.

[7] Y. Takenaka, Y. Kimura, K. Narita, and D. Kaneko, “Mathemat-ical model of direct reduction shaft furnace and its applicationto actual operations of a model plant,” Computers and Chemi-cal Engineering , vol. 10, no. 1, pp. 67–75, 1986.

[8] N. S. Srinivasan, “Reduction of iron oxides by carbon in acirculating fluidized bed reactor,” Powder Technology , vol. 124,no. 1-2, pp. 28–39, 2002.

[9] O. Levenspiel, Chemical Reaction Engineering , John Wiley &Sons, New York, NY, USA, 2nd edition, 1962.




[10] K. Mondal, H. Lorethova, E. Hippo, T. Wiltowski, and S.B. Lalvani, “Reduction of iron oxide in carbon monoxideatmosphere—reaction controlled kinetics,” Fuel Processing Technology , vol. 86, no. 1, pp. 33–47, 2004.

[11] Y. Iguchi andS. Yokomoto, “Kinetics of the reactions in carboncomposite iron ore pellets under various pressures from vac-uum to 0.1 MPa,” ISIJ International , vol. 44, no. 12, pp. 2008–

2017, 2004.[12] G. V. Reklaitis, Introduction to Material and Energy Balances,

John Wiley & Sons, 1st edition, 1983.[13] D. W. Green and R. H. Perry, Perry’s Chemical Engineer’s

Handbook, The McGraw-Hill Companies, 7th edition, 1999.[14] C. C. Furnas, “Heat transfer from a gas stream to a bed of

broken solids,” Industrial and Engineering Chemistry , vol. 22,no. 1, pp. 26–31, 1930.

[15] P. Munster and H. J. Grabke, “Kinetics of the steam reformingof methane with iron, nickel, and iron-nickel alloys as cata-lysts,” Journal of Catalysis, vol. 72, no. 2, pp. 279–287, 1981.

[16] F. Bustamante, R. M. Enick, A. V. Cugini et al., “High-tem-perature kinetics of the homogeneous reverse water-gas shiftreaction,” AIChE Journal , vol. 50, no. 5, pp. 1028–1041, 2004.

[17] J. Shi, E. Donskoi, D. L. S. McElwain, and L. J. Wibberley,“Modelling the reduction of an iron ore-coal composite pelletwith conduction and convection in an axisymmetric tempera-ture field,” Mathematical and Computer Modelling , vol. 42, no.1-2, pp. 45–60, 2005.

[18] Empowerment and Research Society Institution of BandungInstitute of Technology (LPPM ITB), “Final report of directreduction process simulation on Krakatau Steel Plant for pre-implementation of zero reformer process,” p. 55, June 2005.

[19] B. Alamsari, S. Torii, A. Trianto, and Y. Bindar, “Numericalsimulation of iron ore reactor isobaric and cooling zone toinvestigate total carbon formation in sponge iron,” in Pro-ceedings of the International Conference of Modeling and Sim-ulation, pp. 88–92, Tokyo, Japan, May 2009.

[20] C. M. Diaz, A. Vahed, D. Shi, C. D. Doyle, A. E. M. Warner, andD. J. MacVicar, “Low temperature thermal upgrading of later-itic ores,” U S patent no. 5178666, 1993.

Documents

Heat andMass Transfer in Reduction Zone of Sponge Iron Reactor.pdf