numerical study on pulverized biochar injection in blast

1521 © 2014 ISIJ

ISIJ International, Vol. 54 (2014), No. 7, pp. 1521–1529

Numerical Study on Pulverized Biochar Injection in Blast Furnace

Agung Tri WIJAYANTA,1,2) Md. Saiful ALAM,3,4) Koichi NAKASO,3) Jun FUKAI,3)* Kazuya KUNITOMO5) and Masakata SHIMIZU5)

1) Research and Education Center of Carbon Resources, Kyushu University, 6-1 Kasuga-koen, Kasuga, Fukuoka, 816-8580Japan. 2) Department of Mechanical Engineering, Graduate School of Engineering, Sebelas Maret University, Jl. Ir.Sutami 36 A Surakarta, 57126 Indonesia. 3) Department of Chemical Engineering, Graduate School of Engineering,Kyushu University, 744 Motooka, Nishi-ku, Fukuoka, 819-0395 Japan. 4) Department of Petroleum and MiningEngineering, Shahjalal University of Science & Technology, Sylhet, 3114, Bangladesh. 5) Department of MaterialsProcess Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka, 819-0395 Japan.

(Received on November 26, 2013; accepted on March 4, 2014)

The possibility of injecting pulverized biochar instead of conventional pulverized coal in blast furnaceironmaking was investigated numerically. More detailed reactions including the water-related reactionswere considered here. The combustion process from the tuyere to the raceway of a blast furnace wassimulated. Oak char (volatile matter wt.% dry basis, VM = 27.11 wt.%-db) provided a lower temperaturethan Taiheiyo coal (VM = 44.60 wt.%-db). Increasing the O2 concentration from 23 to 27 wt.% resulted ina higher combustibility of both solid fuels. However, the effect of increasing oxygen concentration wasstill insufficient for the Oak char at high injection rates because of its inadequate volatile content. Biocharproperties become increasingly important as the injection rate increases. Compared with Oak char thatprovided a combustibility of 68% at an injection rate of 200 [(kg solid fuel)/(1 000 Nm3 feed gas)] and hotblast of 27 wt.% O2 concentration, Oak char 1 (VM = 32.09 wt.%-db) had a higher combustibility of 71%.

KEY WORDS: numerical simulation; blast furnace ironmaking; biochar; combustibility.

1. Introduction

An increase in the requirements for energy savings andreduced environmental impacts has led to investigations intoinnovative ironmaking technologies with the aim of reduc-ing energy consumption and CO2 emissions. CO2 emissionsin the ironmaking process account for approximately 70%of total CO2 emissions in the steel industry.1) Blast furnaceswill remain the predominant ironmaking equipment in theforeseeable future. In a blast furnace, preheated air and fuel(generally pulverized coal) are injected into the lower partof the furnace through a tuyere, forming a raceway in whichthe injected fuel and some of the coke descending from thetop of the furnace are combusted and gasified. An analysisof the heat and mass transfer phenomena inside a blast fur-nace is important; however, it is almost impossible to mea-sure all of the required information inside the blast furnaceaccurately.

Operation with pulverized coal injection (PCI) into a blastfurnace tuyere is used to reduce the coke feed rate. Theoperating conditions, coal types, pulverized coal (PC) diam-eter and PC injection method influence the combustionbehavior in the raceway zone.2,3) The raceway zone is pri-marily responsible for the production of combustion gasesin the blast furnace. When the PC feed rate increases,

unburned char accumulated in the furnace causes a decreasein the permeability through the coke bed. A stable state ofoperation can therefore not be achieved. The combustionprocess in the blast furnace raceway is clearly complex, anddetailed measurements are extremely difficult to obtainbecause of the existing high temperatures and pressures,presence of molten material, lack of accessibility and inev-itable reduction in production. An alternative to direct mea-surements is the development of a mathematical model ofPC injection to clarify the limit of the PCI rate.4) Mathemat-ical models for analyzing the effect of uncertain factors onthe combustion characteristics of PC in blast furnace havebeen reported.5–7) The simulation has emphasized also onthe effect of the lance configuration7) and the results havesuggested that a coaxial double-lance configuration withcooling gas channel induces a stronger swirling flow thanthe lance configuration without cooling gas channel.

Energy system involving carbon-based vectors canachieve very low CO2 intensity when it uses energy mix ofcarbon positive and carbon negative technologies. Interest-ingly, negative CO2 emission can offset CO2 emission gen-erated by conventional fossil fuel-fired energy systems.8)

Energy from biomass is estimated as a promising candidatefor carbon-neutral or even carbon negative energy systemswith C-based energy vectors. Biomass is a solid fuel withhigh moisture and volatile content. Biomass also has a lowerlatent heat and density compared with coal. Biomass cangenerally be defined as a hydrocarbon fuel which consistsmainly of carbon, hydrogen, oxygen and nitrogen. A low

* Corresponding author: E-mail: [email protected]: FC, fixed carbon; VM, volatile matter; PC, pulverized coal;PCI, pulverized coal injection.DOI: http://dx.doi.org/10.2355/isijinternational.54.1521

© 2014 ISIJ 1522

ISIJ International, Vol. 54 (2014), No. 7

moisture content is the basis of high combustion quality;therefore, the biomass material should be stored out of therain and aerated.9) The carbonization of biomass enrichesthe carbon content and removes oxygen. The resulting bio-char has an increased energy density,10) although its charac-teristic differs for each biomass source and at each pyrolysistemperature.11) Biochar from biomass materials is a renew-able energy carrier which is competitive on energetic basisand also as a viable option for atmospheric carbon seques-tration.12) However, the increased utilization of biochar isstill limited in the iron and steel industry.13–15)

This paper provides a numerical investigation into thecombustion of pulverized biochar injected into a blast fur-nace for ironmaking. The purpose of this study was to inves-tigate the potential of using pulverized biochar injectioninstead of conventional PCI in blast furnaces. Comparedwith our previous study,15) we here use a more detailed reac-tion mechanism and focus on high injection rates. The tem-perature and combustible gas distribution profiles throughthe tuyere and raceway zone are explained. A comparisonof calculated results between pulverized biochar injectionand PCI is provided. The combustibility under variousconditions is discussed. Overall, the findings provide infor-mation to be used when considering the implementation ofpulverized biochar in the ironmaking blast furnace.

2. Numerical Analysis

The numerical investigation focuses on the combustionbehavior from the tuyere (where there is a single lance forfuel and blast injection for hot gas) to the raceway region ofa blast furnace. The tuyere is used to implement a hot blastand inject the solid fuel from the lance (Fig. 1(a)). In theblast furnace, the raceway is surrounded by a packed bed ofcoke. An assumption regarding the raceway is the absenceof any solid particles such as coke. Taiheiyo coal and twobiochars (namely Oak char and Oak char 1) were used inthis numerical simulation. Properties of the Taiheiyo coal,2)

Oak char16) and Oak char 114) including the proximate andultimate analyses for the solid fuels are summarized inTable 1.

2.1. Mathematical ModelGas-particle flow plays a dominant role in multiphase

flow in an ironmaking blast furnace. A comprehensive mod-el based on continuum-(Eulerian) and discrete-(Lagrangian)types describing the hydrodynamics of gas-particle flow asa discrete particle model has been developed. In this model,the gas phase is treated with a Eulerian frame and describedby the steady-state Reynolds-averaged Navier-Stokes equa-tions and the k-ε turbulence model.17) In discrete phase mod-eling, particles of known size distributions and propertiesare injected into the combustion chamber and tracked usinga Lagrangian approach throughout the computationaldomain. Individual particle trajectories are tracked andsolved for using Newton’s second law of motion. The con-servation equations of this model have been described indetail in our previous study15) and are summarized in Table2. Here, this model is applied with the modification of someadditional reactions as described below.

Fig. 1. (a) Illustration of the tuyere and raceway of a blast furnace (refer to [4]) and (b) geometry and computationaldomain.

Table 1. Solid fuel properties.

Taiheiyo coal Oak char Oak char 1

Proximate[wt.%-db]

FC 39.80 55.60 62.85

VM 44.60 27.10 32.09

Ash 15.60 17.30 5.06

Ultimate[wt.%-daf]

C 77.50 78.11 78.56

H 6.50 2.54 17.49

O 14.80 18.75 3.54

N 1.00 0.48 0.41

S 0.20 0.12 0

HHV [MJ/kg] 26.40 23.05 28.16

db: dry basis; daf: dry ash free; HHV: higher heating value


1523 © 2014 ISIJ

The combustion process is composed of the followingstages: inert heating, the devolatilization of solid fuel parti-cles, and the gaseous combustion of volatiles, followed bythe oxidation and gasification of char. Inert heating occursuntil the particle temperature reaches the vaporization tem-perature. When the particle temperature reaches the vapor-ization temperature, devolatilization (R1) commences for acombusting particle. Significant devolatilization is initiatedat approximately 600 K.18) The devolatilization processreleases volatiles (CαHβOγNδ) and char (C(s)).

.............. R1

Devolatilization indicates that volatiles are releasedaccording to the following expression:

................ (10)

where mp, mp,0, and A0 represent the particle mass [kg], ini-tial particle mass [kg] and rate constant [s–1], respectively.fv,0 and fw,0 indicate the fraction of volatiles initially presentin the particle and evaporating material, respectively.

The combustion of volatiles is represented by the gasreactions as follows:

........ R2

......................... R3

...................... R4

..................... R5

.................... R6

......................... R7

The stoichiometric coefficients of reaction R2 are sum-marized in Table 3. The formula of volatiles (CαHβOγNδ) isdefined based on the solid fuel properties in the Table 1.Compared with our previous study15) that only used the twooverall reactions R2 and R3, more detailed reactions includ-ing the water-related reactions (R4–R7) are considered here.

The finite-rate/eddy-dissipation model for the gas reac-tion mechanisms in turbulent flow was employed where thereaction rates of the Arrhenius model, Rg,Arr, and the EddyBreak-Up turbulence chemistry interaction model, Rg,EBU,R

for reactants and Rg,EBU,P for products,19) were calculated.The net reaction rate is taken as the minimum of these threerates. Rg,Arr, Rg,EBU,R and Rg,EBU,P are expressed as follows:

Table 2. Conservation equations.15)

Gas phase:

Mass (1)

Momentum (2)

Energy (3)

Gas species i (4)

Turbulent kinetic energy (5)

Turbulent dissipation rate (6)

Particle phase:

Mass (7)

Momentum (8)

Energy (9)

Note: S is source [units vary]; k and ε indicate turbulent kinetic energy [m2/s2] and turbulent dissipationrate [m2/s3]. Gk represents the generation of turbulence kinetic energy related to the mean velocity gradientand C1ε, C2ε, σk, σε are the turbulent model constants.

∇ ( ) =. ρu Sm

∇ ( ) = −∇ +∇ ( ) + +. .ρ τ ρuu p g F

∇ +( )( ) = −∇ ⎛

⎝⎜⎜

⎞

⎠⎟⎟ +∑. .u H p h J Sj j

jhρ

∇ ( ) = −∇ + +. .ρuY J R Si i i i

∂∂

( ) = ∂∂

+⎛

⎝⎜

⎞

⎠⎟∂∂

⎡

⎣⎢⎢

⎤

⎦⎥⎥+ − − +

xku

x

k

xG Y S

ii

j

t

k jk M kρ μ

μσ

ρε

∂∂

( )= ∂∂

+⎛

⎝⎜

⎞

⎠⎟∂∂

⎡

⎣⎢⎢

⎤

⎦⎥⎥+ − +

xu

x xC

kG C

kS

ii

j

t

jkρε μ

μσ

ε ε ρ ε

εε ε1 2

2

εε

dm

dtmp =

du

dtF u u

gFp

D pp

p

= − +−

+( )( )ρ ρρ

m cdT

dth A T T

dm

dtH A T Ti conv i reac Radp p

pp g p

pp p p

4= −( )+ + −( ), , ε σ 4

Solid Fuel C H O N C s → + ( )α β γ δ

− = −( )dm

dtA f f mw

pp0 0 0 01υ , , ,

C H O N O CO H O N2 2 2α β γ δ + → + +a b c d

CO O CO2 2+ →0 5.

CO H O CO H2 2 2+ ↔ +

CO H CH H O2 4 2+ ↔ +3

Table 3. Stoichiometric coefficients of volatile combustion (R2).

Reaction coefficient Taiheiyo coal Oak char Oak char 1

α 1.44 0.83 0.91

β 3.60 2.30 3.11

γ 0.52 1.07 0.97

δ 0.0399 0.0316 0.0259

a 1.36 0.45 0.75

b 1.44 0.83 0.91

c 1.80 1.15 1.55

d 0.0199 0.0158 0.0129

CH O CO H4 2 2+ → +0 5 2.

H O H O2 2 2+ ↔0 5.

© 2014 ISIJ 1524


..... (11)

.......... (12)

........... (13)

ν, Mw, Y, C, are stoichiometric coefficient, molecular weight,mass fraction and molar concentration [kmol/m3] of the cor-responding species, respectively. α is rate exponent for reac-tants and products. Subscript R and P indicate reactant andproduct of reactions. Subscript f and b are forward and back-ward reactions. ρ represents the density [kg/m3]. A and B arethe empirical parameters. The Arrhenius kinetic rate of reac-tion is defined as:

.......................... (14)

where subscript f and b represent forward and backwardreaction. R is the universal gas constant. Ar, β r and Er are thepre-exponential factor, temperature exponent and activationenergy for reaction, respectively.

For heterogeneous surface reactions, the following reac-tions are considered during the combustion process.

......................... R8

......................... R9

.................... R10

The chemical reaction rate is expressed as follows:20,21)

.................... (15)

Ap and η r indicate the particle surface area [m2] and effec-tiveness factor, respectively. Yj is the mass fraction of spe-cies j. p represents the partial pressure [Pa]. The diffusionrate coefficient, D0,r, for reaction r is computed as:

................. (16)

where Cj,r is the molar concentration of species j in reactionr. dp represents particle diameter [m]. The kinetic parame-ters to determine kr using the Arrhenius expression are sum-marized in Table 4.22)

The change in particle temperature is determined using anenergy balance for particles (Eq. (9)) governed by convec-tive heat transfer and latent heat transfer associated withmass transfer and radiative heat transfer.23,24) cp, Hreac and Tin the Eq. (9) represent heat capacity [J/kg-K], reaction heat[J/kg] and temperature [K], respectively. Subscripts p, g andRad, respectively, indicate particle, gas phase and radiationterm. εp and σ are the particle emissivity and Stefan-Boltzmann constant, respectively. P-1 radiation model isused and heat source due to particle radiation expressed asfollows:

.......... (17)

where Ep is the equivalent emission of the particles, a indi-cates the equivalent absorption coefficient, and n representsthe refractive index of the medium. TRad is calculated usingthe following expression:

........................... (18)

G is the incident radiation [W/m2] expressed as follows:

........................... (19)

where I is the radiation intensity and Ω is the solid angle.hi,conv (Eq. (9)) is associated with the Nusselt number,

which is a function of particle Reynolds number and gasPrandtl number25,26) as follows:

.......... (20)

Pr is the Prandtl number of the continuous phase (= cpμ /kα)and Red is the relative Reynolds number based on the parti-cle diameter and relative velocity as follows:

....................... (21)

where kα is the thermal conductivity [W/m-K]. μ and u indi-cate the molecular viscosity [Pa-s] and velocity [m/s],respectively.

2.2. Solution Procedure and Calculation ConditionsAn iterative solution procedure is used in the overall gas-

particle coupling for solving the governing equationsdescribed above.27) As the particle trajectory is computed,the two-way coupling incorporates the effect of the discretephase trajectories on the continuum and the continuousphase always impacts the discrete phase. The iterative cycleis repeated until overall convergence is reached for bothphases. The discretization of the gas phase governing equa-tions is based on the finite volume method employing astaggered grid and solved by the SIMPLE algorithm28) forpressure-velocity coupling.

g,ArrR M k C k CP R w r f R r b PPR

R P= −( ) −⎛

⎝⎜

⎞

⎠⎟∏∏ν ν α α

, ,

g,EBU,R M Ak

Y

MR R wR

R

R w R

=⎛

⎝⎜⎜

⎞

⎠⎟⎟ν ρ ε

νmin

,

g,EBU,R M ABk

Y

MP P w

PP

P w PP

=∑

∑ν ρ ε

ν ,

k A T er r

ERTr

r

=−

β

C(s) O CO2+ →0 5.

C(s) CO 2CO2+ →

C(s) H O CO H2 2+ → +

R A Y pk D

D kj r r jr r

r r,

,

,

=+pη0

0

D CT T

dr r0 1

0 752

, ,

./

=+( )⎡⎣ ⎤⎦p g

p

Table 4. Kinetic reaction parameters.

Reaction Ar βr Er [J/kmol]

R2 2.1 × 1011 0 2.03 × 108

R3 2.2 × 1012 0 1.67 × 108

R4f 2.75 × 1011 0 8.38 × 108

R4b 2.65 × 10–2 0 3.96 × 103

R5f 5.12 × 1014 0 2.73 × 104

R5b 4.4 × 1011 0 1.26 × 108

R6 3 × 108 –1 1.68 × 108

R7 6.8 × 1015 0 1.26 × 108

R8 1.36 × 106 0.68 1.30 × 108

R9 6.78 × 104 0.73 1.63 × 108

R10 8.55 × 104 0.84 1.40 × 108

f: forward reaction; b: backward reaction

−∇ = −⎛

⎝⎜

⎞

⎠⎟ + +( ).q an

Ta a Gr p4 2

4

π σπ

TG

Rad =⎛⎝⎜

⎞⎠⎟4

14

σ

G I d= ΩΩ=∫ 4π

Nuh d

kii conv i

d= = +, . . Re Prp 12

α

2 0 0 61

3

Redd u u

=−ρ

μg p p g


1525 © 2014 ISIJ

2.3. Calculated ConditionsA 2D simulation was used for the tuyere (consisting of

lance and hot blast) and raceway regions of the blast furnace.Figure 1(b) shows the geometry for the computational domain.A fine mesh was developed with 76 646 cells. The numericalinvestigation focused on simulating the tuyere (where thereare single lances for injecting solid fuel and a blast injectionfor hot gas) and the raceway region of the blast furnace. Thelance had an inner diameter of 14 mm. The tuyere had an innerdiameter of 180 mm at the inlet side and 140 mm at its end(Fig. 1(b)). The wall along the tuyere is maintained at adiabat-ic conditions of zero heat flux. The raceway is modeled as atube 0.5 m in diameter and 1.5 m in length. In the blast fur-nace, the raceway is surrounded by a packed bed of coke andis assumed to be at 2 073 K.2) A particle diameter of 70 μmhas been used generally. The computational conditions inTable 5 were selected for investigating the effect of injectionrate and O2 concentration on combustibility.

3. Results and Discussion

A comparison between the results from the modeldescribed above and experimental data29) using a drop tubereactor is shown in Fig. 2. The electrically-heated drop tubefurnace had a length of 2.5 m and an internal diameter of200 mm. Bituminous coal was injected at a rate of 1 kg/hwith carrying air of 1.5 Nm3/h at 473 K and total gasifying airof 8 Nm3/h at 523–623 K. The wall of the reactor made fromceramic was maintained at 1 523 K. A level of confidence inthe predicted O2 and CO2 was established through comparisonwith experimental results. It is expected that the simulationmethod is capable of providing realistic predictions.

The Oak char provided a lower temperature distributionthan the Taiheiyo coal at an injection rate of 200 [(kg solidfuel)/(1 000 Nm3 feed gas)] and hot blast of 23 wt.% O2

(Fig. 3(a)). The field temperature of the Oak char was lowerthan that of the Taiheiyo coal. This result is expectedbecause the combustion heat of Oak char is lower than thatof Taiheiyo coal. The temperature distribution influences thegas composition (Fig. 3(b)). As shown in Fig. 3(b), theinjection of Oak char with low volatile content causes a low-er overall burnout. The volatile content results in the firstignition in the tuyere. Volatile contents are released fromsolid fuels during devolatilization. Solid fuels with high vol-atile contents are normally injected because of their gener-ally superior combustion performance resulting from betterC(s) reactivity and hence burnout. Because of volatile oxi-dation (R2), the reduced release of volatile content by theOak char compared with that of the Taiheiyo coal results in

the Oak char having a higher oxygen composition than theTaiheiyo coal (Fig. 3(b)).

The released volatiles burn rapidly with oxygen in the hotblast thereby increasing the gas and particle temperatures.Both the average gas and particle temperatures of Oak charwere lower than the Taiheiyo coal at an injection rate of 200[(kg solid fuel)/(1 000 Nm3 feed gas)] and a hot blast of 23wt.% O2 (Fig. 4(a)). A decrease in average particle temper-ature occurred at an axial distance approximately 0.2 m intothe tuyere because of the low carrier gas temperature. Thetemperature history of the particles influences the local con-centrations of gas species (Fig. 4(b)). Compared with theTaiheiyo coal, the Oak char cannot yield a higher volatilecombustion (R2) because it contains an inadequate volatilecontent. As a result, the Oak char yielded a higher oxygenconcentration than the Taiheiyo coal. Compositions of COand CO2 are influenced primarily by reaction R3. The relat-ed water reactions, mainly R7, favor water production.

Figures 5(a) and 5(b) show the effect of O2 concentrationon average temperature and gas compositions along the axisat an injection rate of 200 [(kg solid fuel)/(1 000 Nm3 feedgas)] for the Taiheiyo coal and Oak char, respectively. Forboth samples, using a hot blast with 27 wt.% O2 increasesthe temperature and influences gas concentrations. Thisoccurs because O2 enrichment improves the combustionefficiency. This additional O2 portion from the hot blastaffects the temperature history of the particle and can influ-ence the local concentrations of gaseous species. Anincrease in O2 concentration increases the temperature andCO2 content because of the exothermic reactions C(s) +0.5O2 → CO (R8) and CO + 0.5O2 → CO2 (R3). Further-more, oxygen in the raceway is consumed rapidly as the pul-verized solid fuel is injected. The maximum CO2 concentra-

Table 5. Computational conditions.

Hot Blast:Inlet velocity = 188 [m/s] O2 composition = 23, 27 [wt.%]

Temperature = 1 450 [K]

Solid Fuel:

Feed rate = 25 – 200 [(kg solid fuel)/(1 000 Nm3 feed gas)] *

Temperature = 300 [K]

Particle diameter = 70 [μm]

Carrier:Inlet velocity = 1.13 [m/s] O2 composition = 23 [wt.%]

Temperature = 300 [K]

* Feed gas = Hot blast + Carrier

Fig. 2. Comparison between experimental29) and calculated results.

© 2014 ISIJ 1526


tion corresponds to that of highest temperature. Due to theexothermic reaction of C(s) + 0.5O2 → CO (R8), the highertemperature at the end part of raceway occurs (see also Fig.3(a)).

As shown in Figs. 5(a) and 5(b), the effect of oxygenenrichment for the Oak char was smaller than that for theTaiheiyo coal. An increase in oxygen concentration in thehot blast promotes volatile combustion (R2). This causes anincreased concentration of combustible gas CO that resultsin an increased temperature, as reported in our previousstudy.30,31) However, the effect of oxygen concentrationdecreases at an axial distance of 1.8 m for the Taiheiyo coal(Fig. 5(a)) at the 27 wt.% O2 concentration. In this position,the average temperature decreased and became lower thanthat at 23 wt.% O2 concentration. The decrease in averagetemperature and increase in oxygen concentration resultfrom the completion of volatile oxidation (R2). A decreasein CO2 concentration occurs at this position because there isinsufficient CO for CO oxidation to proceed (R3). As for theTaiheiyo coal, the average temperature at 27 wt.% O2 con-centration decreased and became lower than that at 23 wt.%O2 concentration at almost 2 m for the Oak char (Fig. 5(b)).For both the Taiheiyo coal and the Oak char, the effect ofO2 enrichment is limited at high injection rates of 200 [(kgsolid fuel)/(1 000 Nm3 feed gas)]. A limitation of O2 enrich-ment is more apparent for the Oak char compared with theTaiheiyo coal. The Oak char cannot sustain further volatilecombustion at high O2 concentration because it containsinsufficient volatile matter. The volatile content thereforebecomes increasingly important at high injection rates.

Fig. 3. (a) Temperature profile and (b) gas compositions at 200 [(kg solid fuel)/(1 000 Nm3 feed gas)] and 23 wt.% O2.(Online version in color.)

Fig. 4. Comparison of average (a) temperature and (b) gas composi-tions for Taiheiyo coal and Oak char at 200 [(kg solid fuel)/(1 000 Nm3 feed gas)] and 23 wt.% O2 in the axial direction.


1527 © 2014 ISIJ

Figures 6(a) and 6(b) show the effect of O2 concentrationon the reaction rate of volatile combustion R2 at an injectionrate of 200 [(kg solid fuel)/(1 000 Nm3 feed gas)] for theTaiheiyo coal and Oak char, respectively. These figures pro-vide added information regarding the limitation of oxygenenrichment and illustrate why the average temperature at 23wt.% O2 was higher than that at 27 wt.% O2 at the end ofthe raceway zone. An increase in temperature occurs late at23 wt.% O2 because the exothermic C(s) oxidations aredelayed by the lag in volatile oxidation. It also causes theparticle temperature of Oak char lower than that of Taiheiyocoal (see Fig. 4(a)). Compared with the Taiheiyo coal in Fig.6(a), the Oak char (Fig. 6(b)) has a low volatile oxidationrate and a delay in volatile oxidation for both the 23 and 27wt.% O2 concentrations.

Figure 7 shows the variation in combustibility with injec-tion rates under various conditions. The combustibility wasdetermined using the C(s) mass fraction loss (mout) at thetuyere exit and the C(s) mass fraction from the original solidfuel (min) at the entrance of the lance:

...................... (22)

Figure 7(a) presents a variation in combustibility with hotblast at 23 and 27 wt.% O2 for the Oak char and Taiheiyocoal. Oak char with a lower volatile content tended to pro-vide a lower combustibility. At 27 wt.% O2, the combusti-bility for both Oak char and Taiheiyo increased because theconsumption oxygen increases for reaction with unburntC(s). While Oak char appears to have little significantimpact on blast furnace operation at a low injection rate,biochar properties become increasingly important at highinjection rates. With higher volatile content, a higher oxygenconcentration in the hot blast provides an increased combus-tibility at a high injection rate.

Figure 7(b) compares the combustibility between themodel with and without the water reactions for Oak char at200 [(kg solid fuel)/(1 000 Nm3 feed gas)]. The reactionsR4–R7 were neglected in the no-water reaction model.Compared with the model without water reactions, the com-bustibility of the model including water reactions offered abetter result for both the 23 and 27 wt.% O2. The modelincluding the additional water related reactions provided a

Fig. 5. Effect of O2 concentration at 200 [(kg solid fuel)/(1 000Nm3 feed gas)] on average temperature and gas composi-tions for (a) Taiheiyo coal and (b) Oak char. (Online ver-sion in color.)

Fig. 6. Effect of O2 concentration at 200 [(kg solid fuel)/(1 000Nm3 feed gas)] on reaction rate of volatile combustion R2for (a) Taiheiyo coal and (b) Oak char. (Online version incolor.)

η =−

×m m

min out

in

100%

© 2014 ISIJ 1528


Fig. 7. Combustibility with injection rates under various conditions: (a) using the model, (b) comparison between modelsfor Oak char with and without water reactions, (c) comparison of Oak char for models with and without radiationand (d) comparison between Oak char and Oak char 1 using the model. (Online version in color.)


1529 © 2014 ISIJ

higher gas temperature distribution in the raceway than thatwithout because of the significant effect of the exothermicreactions from the oxidation reactions of R6 and R7.

Figure 7(c) explains the combustibility between the no-radiation and radiation models for Oak char at 200 [(kg solidfuel)/(1 000 Nm3 feed gas)]. The third term on the right handside of Eq. (9) was not considered in the no-radiation model.Consequently, the no-radiation model did not include theradiation heat flux (Eq. (17)) in the heat sources of the ener-gy equations. The model including radiative heat transfershowed a significant effect on combustibility for both thehot blasts of 23 and 27 wt.% O2. This is because the energyequation accounts for radiative heat transfer and increasesthe particle temperature history thereby increasing the gastemperature distribution. The radiative heat transfer there-fore becomes important to account for the prediction.

Figure 7(d) compares the combustibility between Oakchar (VM = 27.11 wt.%-db) presented on the red coloredline and Oak char 1 (VM = 32.09 wt.%-db) on the blackline. Oak char 1 with higher volatile content and calorificheating value (see Table 1) provided a higher combustibility.At high injection rates of 200 [(kg solid fuel)/(1 000 Nm3

feed gas)] and a hot blast of 27 wt.% O2 concentration, Oakchar 1 offered a higher combustibility of 71% comparedwith Oak char that provided the combustibility of 68%.Biochar properties become increasingly important at highinjection rates (also see Fig. 7(a)).

4. Conclusions

The possibility of pulverized biochar injection in blastfurnace ironmaking was numerically investigated. In thepresent model, heat and mass transfer with more detailedreactions including the water-related reactions for the vola-tile matter as well as the devolatilization of the solid fuelparticle were taken into account. Pulverized Oak char wasespecially studied as a candidate of the solid fuel to the blastfurnace. Its combustion behavior in the tuyere and racewayof the blast furnace was studied and compared with conven-tional pulverized Taiheiyo coal injection. The major resultsand findings were listed below:

(1) The present model reasonably agreed with theexperimental data from the literature.

(2) Temperatures of gas and particle in the tuyere andraceway with Oak char were lower than those with Taiheiyocoal by several hundred degrees. Lower concentration ofCO2 and higher concentrations of O2 and CO were obtainedfor the case of Oak char compared with the case of Taiheiyocoal. These results were attributed to the lower content ofvolatile matter of Oak char than that of Taiheiyo coal. Thecontent of volatile matter played an important role for thegas temperatures and the gas concentrations in the tuyereand raceway.

(3) The effects of increases in O2 concentration on thetemperatures and the gas concentrations were limited for thecase of Oak char compared with Taiheiyo coal. Oak charcould not sustain further volatile combustion at high O2

concentration because of its insufficient content of volatilematter. The content of volatile matter therefore became sig-nificant at high injection rates.

(4) According to the calculation results using threekinds of solid fuel, the combustibility of the solid fuelincreased with its content of volatile matter.

(5) The radiation heat transfer strongly contributed tothe temperature distributions in the tuyere and raceway andthe combustibility of the solid fuel. Therefore, not only theproperties of the radiation heat transfer such as the coeffi-cient of adsorption but also concentrations of CO2 and H2Owere important for the prediction.

(6) The reactions concerning water significantly influ-enced the temperature distribution in the tuyere and racewayand the combustibility of the solid fuel because of the exo-thermic oxidation reactions as well as the radiation heattransfer.

These results and findings can contribute to an under-standing of the pulverized biochar injection with the aim ofachieving low emission blast furnace ironmaking.

AcknowledgmentsThis research work was partially supported by the Japan

Society for the Promotion of Science (JSPS) ScientificResearch (A), 2010–2011, Research Number 22241020. Theauthors also gratefully acknowledge a grant from the GlobalCenter of Excellence in Novel Carbon Resource Sciences,Kyushu University.

REFERENCES

1) Y. Ueki, R. Yoshiie, I. Naruse, K. Ohno, T. Maeda, K. Nishioka andM. Shimizu: Fuel, 104 (2013), 58.

2) H. Aoki, H. Nogami, H. Tsuge, T. Miura and T. Furukawa: ISIJ Int.,33 (1993), 646.

3) H. Nogami, H. Yamaoka and K. Takatani: ISIJ Int., 44 (2004), 2150.4) L. Takeda and F. C. Lockwood: ISIJ Int., 37 (1997), 432.5) K. Nishioka, T. Maeda and M. Shimizu: ISIJ Int., 45 (2005), 669.6) S. Matsuzaki, A. Shinotake, M. Naito, T. Nishimura, K. Kunitomo

and T. Sugiyama: Nippon Steel Tech. Rep., 94 (2006), 87.7) C. P. Yeh, S. W. Du, C. H. Tsai and R. J. Yang: Energy, 42 (2012), 233.8) W. M. Budzianowski: Energy, 41 (2012), 280.9) A. Strehler: Ecol. Eng., 16 (2000), S25.

10) H. Abdullah and H. Wu: Energ. Fuel., 23 (2009), 4174.11) S. W. Park and C. H. Jang: Energy, 39 (2012), 187.12) D. Matovic: Energy, 36 (2011), 2011.13) H. Helle, M. Helle, H. Saxén and F. Pettersson: ISIJ Int., 49 (2009), 1316.14) A. Babich, D. Senk and M. Fernandez: ISIJ Int., 50 (2010), 81.15) A. T. Wijayanta, M. S. Alam, K. Nakaso, J. Fukai, K. Kunitomo and

M. Shimizu: Fuel Proc. Technol., 117 (2014), 53.16) S. Gaur: Thermal Data for Natural and Synthetic Fuels, Marcel

Decker, New York, (1998).17) B. E. Launder and D. B. Spalding: Mathematical Models of Turbulence,

Academic Press, London, (1972).18) M. M. Baum and P. J. Street: Combust. Sci. Tech., 3 (1971), 231.19) B. F. Magnussen and B. H. Hjertager: Proc. of 16th Int. Symp. on

Combust., The Combustion Institute, Pittsburgh, PA, (1976), 719.20) I. W. Smith: Proc. of 19th Int. Symp. on Combust., Combust. Insti-

tute, Pittsburgh, PA, (1982), 1045.21) A. Silaen and T. Wang: Int. J. Heat Mass Transf., 53 (2010), 2074.22) H. Watanabe and M. Otaka: Fuel, 85 (2006), 1935.23) R. Viskanta and M. P. Mengüç: Prog. Energy Combust. Sci., 13

(1987), 97.24) M. Hossain, J. C. Jones and W. Malalasekera: Flow Turbul. Com-

bust., 67 (2001), 217.25) W. E. Ranz and W. R. Marshall, Jr.: Chem. Eng. Prog., 48 (1952), 141.26) W. E. Ranz and W. R. Marshall, Jr.: Chem. Eng. Prog., 48 (1952), 173.27) Ansys Fluent 12.0, User’s Guide and Database, Ansys Inc., Canonsburg,

(2009).28) S. V. Patankar: Numerical Heat Transfer and Fluid Flow, Hemisphere,

Washington DC, (1980).29) Y. Zhang, X. L. Wei, L. X. Zhou and H. Z. Sheng: Fuel, 84 (2005), 1798.30) A. T. Wijayanta, M. S. Alam, K. Nakaso and J. Fukai: Fuel, 93

(2012), 670.31) A. T. Wijayanta, M. S. Alam, K. Nakaso, J. Fukai and M. Shimizu:

Bioresource Technol., 110 (2012), 645.

numerical study on pulverized biochar injection in blast

Documents