Operation and Simulation of Pressurized Shaft Furnace for ...

Operation and Simulation of Pressurized Shaft Furnace

for Direct Reduction*

By Reijiro TAKAHASHI, * * Yoshikazu TAKAHASHI, * * * Jun-ichiro and Yasuo OMORI * *

YAGI **

Synopsis A laboratory scale shaft furnace was constructed for examining the op-

erational characteristics. Some series of reduction experiments of i ron oxide

pellets with gas mixture of H2, CO and others were carried out under various operational conditions varying temperature, flow rate and composi-tion of inlet gas and pressure applied. Overall data of operation and dis-tributions of process variables were obtained for the steady-state operations. After cooling down the furnace, the pellets were sampled from various levels in the furnace. Then, the measurement of crushing strength and the ob-servation of reduction fashion of the pellets were conducted.

On the other hand, a one dimensional mathematical model was derived

for the process simulation of the shaft furnace. The rate parameters for the reduction of the pellets and the side reactions which were measured by independent experiments were incorporated in the model. Reasonable simulation results were obtained for ordinary pressure operation when ap-

propriate values of the rate parameters were given for the reduction of iron oxide pellets and side reactions.

I. Introduction

In recent years, the direct reduction process with

reducing gas has attracted special interest and the

process has already been applied practically from the viewpoint of effective use of various kinds of energy

sources for providing stable supply of sponge iron instead of high quality scrap, and of improving the environment in the ironmaking process. It is esti-

mated that the reduced iron will be mainly produced

by shaft furnace process, particularly in high pressure operation. The several experiments1-5) with theoretical analy-

sis have been conducted on the reduction of iron

oxide in the shaft furnace at an atmospheric pressure. Although it has been pointed out that the high pres-

sure operation such as NSC direct reduction processs~

and HyL-III process') has some advantages to improve the production rate, few attempts have been made on

the fundamental studies of high pressure reduction in the shaft furnace.

In view of the present status of the direct reduction

mentioned above, a pressurized shaft furnace in labo-ratory scale was constructed by the authors. It is

13 cm in internal diameter and 2 m in effective height. Some series of reduction experiments of iron oxide

pellets with gas mixture of H2, CO and others were carried outa> by using the shaft furnace under the different operational conditions. Among them the

temperature, flow rate and composition of inlet gas and pressure applied were changed thereby examining

the effect of such operational conditions on the

operational results, distribution of process variables, reduction fashion and strength of the pellets.

On the other hand, a one-dimensional mathe-

matical model developed considering both heat and material balances. Process simulation of the shaft

furnace was carried out9~ by using the model which includes kinetics of the reduction of iron oxide pellet

by gas mixture and of side reactions. The model construction and values of the various parameters

included were examined by comparing the calculated values and observed data in the shaft furnace.

II. Experimental Apparatus

Figure 1 shows a schematic diagram of the pres-surized shaft furnace used and Table 1 represents main specifications of the furnace.

The reaction tube made of super heat resistance cast alloy (27Cr-50Ni-W) was mantled with thermal insulator of 10 cm in thickness to keep the heat loss through the wall of the reaction tube as low as pos-

Fig . 1. Schematic diagram

tUS.4~5)

of the experimental appara-

*

**

***

This article was presented to the 43rd Ironmaking Conference, April 1984, in Chicago, by support of the Hyuga Fund of ISIJ. Preprinted with permission from Ironmaking Proceedings, Iron and Steel Society of RIME, PA, 43 (1984), 485. Manuscript received on

January 16, 1986; accepted in the final form on April 18, 1986. © 1986 ISIJ Research Institute of Mineral Dressing and Metallurgy, Tohoku University, Katahira, Sendai 980. Professor Emeritus, Tohoku University; Tetsugen Co., Ltd., Fujimi, Chiyoda-ku, Tokyo 102.

Research Article ( 765 )

(766) Transactions ISI1, Vol. 26, 1986

sible. The inner diameter of the tube was determined to be ten times as long as the diameter of the pellets in order to decrease the wall effect on the respective flow of gas and solids.

Reducing gas which was preheated to a specified temperature was introduced into the reaction tube and water vapor was also given by supplying water into the heat exchanger. The gas entrance consisted of double coaxial circular reaction tube and the inner one had 40 holes (5 mm ID) for circumferentially uniform introduction of gas at the fixed level.

Temperature, pressure and gas composition were automatically measured by sheathed chromel-alumel thermocouples, an aqueous manometer and a process mass analyser (MGA-1200) respectively at seven

points in the longitudinal direction of the furnace. The experiment can be carried out within the limits of maximum gas temperature of 1 373 K, pressure of 0.92 MPa and gas flow rate of 0.05 m3 (STP)/s by using the apparatus.

III. Samples and Experimental Procedure Two kinds of fired iron oxide pellets were used in

the experiments. Chemical composition of the pel-lets is shown in Table 2. W pellet has 3.8 g/cm3 and N pellet has 3.6 g/cm3 as the average value of

apparent density. The pellets were sieved to obtain

particles having the diameter, 1.3 ± 0.1 cm for the shaft furnace experiment.

The purpose of these experiments was to measure the operational results and the distribution of process variables such as temperature, pressure and gas com-

position under the steady-state in the shaft furnace. In addition, the reduction fashion and strength of the

pellets during the reduction were also examined. Main experimental conditions at the entrance in

the furnace are shown in Table 3. The reducing

gas ratios (G/ W = 2.3 N 3.1 m3 (STP)/kg (Fe)) were relatively unchanged except for Runs 14, while effects of reducing gas composition such as concentrations of CO, CO2 and CH4, of the temperature (Runs 7 to 9) and of the pressure of blowing gas (Runs 3 to 6 ) on the distribution of process variables were examined in these experiments. W pellets in Table 2 were only used for Runs 1 and 2, while N pellets were used for the other Runs.

Temperature, pressure and gas composition were measured at seven points in the longitudinal direc-tion of the shaft furnace until reaching the steady-state. After finishing all the measurements, the furnace was cooled down by N2 gas. Then, the pellets were sampled at specified levels of the furnace, reduc-tion degree of the sampled pellets were determined by reducing the residual oxide with hydrogen at 1 323 K.

Iv. Derivation of the Mathematical Model and Calculation Method

One dimensional mathematical models on the

Table 1. Main specification

furnace.

of the pressurized shaft

Table 2. Chemical composition of iron oxide pellets.

Table 3. Experimental conditions and concentration of blowing gas for moving bed reactor.

Research Article

Transactions 'SIT, Vol. 26, 1986 (767)

reduction of iron oxide pellet by the shaft furnace were proposed by the authors5"2~ and other research-ers.s~ In this report, a similar model was derived by taking the heat and material balances under the fol-lowing assumptions:

(1) Steady state. (2) Piston flows of gas and pellets.

(3) Reduction rate of the individual pellet was represented by the three-interface model.10~

(4) Linear addition was allowed for the rates of hydrogen and carbon monoxide reduction of the

pellets. (5) Heat of reaction for reduction of the pellets

was given to the solid flow. (6) Uniform temperature distribution in a single

pellet. (7) Water gas shift reaction and both reactions of

methane formation and decomposition were con-sidered as side reactions.

(8) Heat of reaction for each side reaction was given to the gas.

By taking the material balance, fundamental equations on gaseous concentration and reduction degree of the pellets are represented by Eqs. (1) and

(2): dY/dz = dY(H2)/dz+dY(co)/dz ...............(1)

dRi/dz = dRix2)/dz+dRicol/dz ...............(2)

3 where, dY(k'ldz = (6(1-~B)l ~rdp)(S/G) o2k~ .........(3) a=1

dRZk~ /dz = (6(1 _ EB)l~rd p) (SI W • dov)~ik~ ......... (4)

where, i=1, 2, 3 and 4. ` i ' denotes each reduction step such as Fe

203 - * Fe304, Fe304 --> FexO, FexO -~ Fe and Fe304 -~ Fe, respectively, while k denotes hydrogen or carbonn monoxide reduction by H2 or CO.

By taking the heat balance, fundamental equations on the temperatures of gas and solids are represented by Eqs. (5) and (6) in which m and n indicate the number of solid and gaseous components respectively:

d T gl dz = (6(1 _ SB)l dp)hPS( T g- T S)

+2rD1U(Tg-Ta)

n

x{~ (Gj(cgj+Tg•dCg;/dTg))}-1 ...............(5) j=1

dTS/dz = (6(1 _EB)/dp)S(hp(Tq~ TS)

3

-~ ((4Hz)v2)l(ndp)2)

m

x{~ (Wk(cj~k+TS•dcpk/dTS))}-1 ............(6) k=1

Overall reduction degree may be calculated from the reduction degree of each reduction step by the following expressions:

R7 = 0.111R1+0.189R2+0.700R3 (Tg>848 K)

...........................(7)

RT = 0.111 R1 + 0.889R4 (Tg < 848 K) ............ (8)

Pressure drop in the shaft furnace was estimated

from Ergun's equation as described in the following expression :

4P = (F•GM.u/dp)((1-~B)/rB)(dz/gC•1033.0) ........................... (9)

F =150.0(1-SB)1Rep+1.75 ........................(10)

As the constants in these equations, the voidage of 0.44 which was measured by using the present furnace under the movement of the bed at room temperature was applied to Eq. (9). The value of 6.5 x 10-5

(cal/cm2 • s . K) which was measured5> in the fixed bed of iron oxide pellets, was applied as the overall heat transfer coefficient (U) in Eq. (5) for evaluating the heat loss through the furnace wall. Equations (1) to (10) were solved numerically under following boundary conditions :

z=0: RT=0 and TS= Ts ..... (11) z=L: Y,z=Yz and Tg=Ti n g

Numerical calculation for the model was carried out according to Runge-Kutta method in the direc-tion from the top to the bottom of the shaft furnace. Observed data at the level of 175 cm were adopted as boundary conditions at the top because of the reliable experimental data.

V. Rate Equations and Rate Parameters for Each Reaction

1. Reduction of a Single Iron Oxide Pellet

The multi-interface reaction model proposed by Hara et al.10~ was used as the reduction rate of a single iron oxide pellet. The rate parameter obtained by the data fitting method were applied mainly to the model. For the better process simulation of the shaft furnace as previously reported,11) the rate parameters obtained by a step-wise reduction method were used in the reduction step from Fe304 to FexO in the tem-perature range under 973 K. Furthermore, intra-particle diffusivity for the reduction step from FexO to Fe in the range of the reduction degree over 75 % was modified to one tenth of the observed value because the retardation of the reduction occurred due to the change of the reduction mechanism in the range of high reduction degree. This modification

gave the excellent simulation result of the laboratory-scale shaft furnace as described later in this paper.

Tables 4 and 5 show the temperature dependences of the rate parameters for H2 reduction and for CO reduction, respectively. In addition, temperature dependence of equilibrium constant for respective reduction step described previously'8 was used. Whereas heat of reaction for each step was evaluated from the equation rearranged by Omori et a1.12~ and Hara et al.13~

2. Water Gas Shift Reaction

Water gas shift reaction are given by Eq. (12) :

CO+H20 = C02+H2 ..................(12)

Research Article

(768) Transactions Is", Vol. 26, 1986

The rates of the forward and the reverse reaction were conventionally represented by the equations described in Eqs. (13) and (14) :

rCO2 = VpkWP2(YCOYH2o-YCOZYH2/KW) .........(13)

rCO2 = V pkWP2(YCO2YH2-KWYCOYH2o) .........(14)

The values of the reaction rate constants (kW, kw) included in Eqs. (13) and (14) were determined by the authors in the previous reports14) under the follow-ing constraints. The values of kW and kW which were obtained under the existence of FexO catalyst were used in the reduction degree less than 50 % in the shaft furnace, and in the reduction degree over 50 %, kW and kw for Fe catalyst applied. Temperature dependences of kW and Ic used were shown in Table 6.

Equilibrium constant for the reaction was calcu-lated from the following equation derived by rear-ranging the thermodynamic data15~ :

KW = exp (3 863.7/ Tg-3.5414) ............(15)

Furthermore, heat of reaction for the reaction was evaluated by the following equation reported by Hara et a1.13! :

4HW = -10 000.0+2.31(Tg-273.0) .........(16)

3. Reactions for Formation and Decomposition of Methane

Methane formation reaction is given by Eq. (17) :

CO+3H2= CH4+H2O ..................(17)

The rate equations of the forward and reverse reactions given by Eq. (17) are represented by Eqs.

(18) and (19), respectively:

TCH4 = VpkMP4(YcoYHZ YcH4YH2o/KM) .........(18) TCH4 = V pkMP2(YCH4YH2o-KMYcoYH2) .........(19)

The values of the reaction rate constants (kM, kM) listed in Table 6 were used under the following con-straints. The values of kM and kM obtained under the existence of Fe catalyst were used only over the reduction degree of 50 % in the shaft furnace. How-ever, because the change of CH4 was hardly observed under the oxide catalyst as reported previously,l4j the reactions for formation and decomposition of CH4 was neglected in the reduction degree less than 50 %. Temperature dependences of kM and kM used were presented in Table 6. Furthermore, equilib-rium constant and heat of reaction for the reactions were calculated by Eqs. (20) and (21) derived from the thermodynamic data15~ :

KM = exp (23 700/ l g-8.4241og Tg- 1.571) ........................ (20)

4HM = -45 200-16.59 Tg+9.64 x 10.3 Tg

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

VI. Experimental Results

1. Effect of the Experimental Conditions on the Distribu- tion of Process Variables

The main results obtained from the experiments are shown in Table 7, RF and dP show the final reduc-tion degree of the pellet and the pressure drop in the furnace. ~1co, ~7H2 and ~co+H2 represent the gas utili-zation degrees of carbon monoxide, hydrogen and gas mixture calculated from Eqs. (22) to (24). However,

7CO+H2 is calculated from Eq. (25) applying both reduction degree of the pellets and experimental conditions:

rlco = ((Yc02-Yc02)/Yco) x 100 ........................(22)

7H, - ((YH,0 YH2o)/YH2) x 100 ........................(23)

7CO+H2 = {((Yc02+Y0)-(Yi;02+YH2O))/(Yc0+YH2){

x 100 .............................................(24)

r1CO+H2 = ((W+Ox xRF)/(16 x G) x 100 ......(25)

1. Effect of Concentration of Carbon Monoxide In order to examine the effect of CO concentra-

tion on the distribution of process variables in the

Table 4. Temperature dependences of the rate param-

eters obtained by data fitting method with

hydrogen reduction.9>

Table 6. Temperature dependences

rate constants obtained by

for side reactions.14~

of

the

the reaction

experiments

Table 5. Temperature dependences of the rate param-

eters obtained by data fitting method with

CO-C02 reduction.9~

Research Article

Transactions ISIJ, Vol. 26, 1986 (769)

furnace, Runs 1 to 3 were carried out with different CO concentration at ordinary pressure. The results obtained are shown in Fig. 2, W pellet was used for Runs 1 and 2 with slightly different temperature of blowing gas, while N pellet for Run 3 in Fig. 2 was used as a reference (see Fig. 3).

According to the distribution of reduction degree by hydrogen reduction (Run 1) in Fig. 2, the stagna-

tion of the reduction occurred at the reduction degree of around 10 N 15 % in the upper part of the bed. The reason was that the operating line in the opera-tion diagram16> for the shaft furnace approached closely to M point (Equilibrium in Fe3O4 FexO-H2O-H2 system). The phenomena may be considered as a special feature in the hydrogen reduction.5~ On the other hand, in the gas mixture reduction (Runs 2

Table 7. Experimental results of the moving bed.

Fig. 2. Comparison of longitudinal distributions of

process variables between the hydrogen re-duction (Run 1) and the reduction with

gas mixture (Runs 2 and 3).

Fig. 3.

Effect of the blowing gas

pressure on the longitu-

dinal distributions of

process variables.4~

Research Article

(770) Transactions ISIT, Vol. 26, 1986

and 3), the higher distribution of temperature was observed with increase of CO concentration. Con-sequently, the distribution of reduction degree was also shifted upwards, thereby improving well the operational results in the shaft furnace.

The accelerative effect of the reduction rate due to CO addition had been observed and explained by the exothermic reactions of the CO reduction by Hara et a1.13~ and Narita et a1.17~ However, in the authors' view based on the fundamental experiments for the reduction rate of a single iron oxide pellets11 and side reaction,14~ these phenomena can be explained as follows. Table 8 shows the heat of reaction for each reduc-tion step and of side reactions. The effect of the CO reduction was understood not to be significant by considering that the rate parameter of each reduction step for CO reduction was in the magnitude from 7 to 14 % of those for H2 reduction9~ and CO concen-tration was remarkably smaller than H2 concentra-tion in this experimental conditions. On the other hand, water gas shift reaction has the heat of reaction of 9 940 cal/mol shown in Table 8, furthermore, the reaction rate which was measured by the authors14~ was considerably large because iron oxide acted as the catalyst. From the fact mentioned above, it was concluded that the better results obtained in the gas mixture reduction were due to the duplicated effects of increase of H2 reduction potential at the reaction interface and the exothermic phenomenon by the water gas shift reaction. 2. Effect of Blowing Gas Pressure

Figure 3 shows the results obtained from Runs 3 to 6 in order to examine the effect of blowing gas

pressure on the distribution of process variables. In Table 7, no values of ~lco+x2/rlco+x2 exceeds

1.0 because the apparent gas utilization degree, calcu-lated from the values of gas analysis was low according to the side reaction in the furnace. It was also found that the pressure drop, 4P decreased with increasing the pressure.

In Fig. 3, the concentration of oxidizing gas and of CH4 in Run 4 (Pi=0.304 MPa) were higher than those in Run 3 (Pi=0.152 MPa). CH4 which was

produced in the heat exchanger for preheating the gas decomposed rapidly in the entrance region of the shaft furnace, but in the rest part of the furnace it was almost unchanged. In the upper part of the furnace, the reduced iron was not yet produced

significantly, in other words, much amount of the unreduced iron oxide was remained. As known alreadyl4) the rates of formation and decomposition reactions of CH4 were considerably low even under the existence of the catalyst. Therefore, no change of CH4 could be found in the furnace. Distribution of reduction degree in Run 5 was shifted upwards in comparison with that of Run 3; however, higher reduction degree of discharged pellets was not obtained in Runs 5 and 6 due to high concentration of oxidizing

gas and less effect of pressure increase on the reduc-tion rate of the pellets. 3. Effect of Blowing Gas Temperature

Figure 4 shows relations of final reduction degree of the pellet (RF) and of gas utilization degree ('2) with blowing gas temperature (TT,) obtained from the observed data in Runs 7 to 9 for examining the effect of blowing gas temperature on the operational results.

In Fig. 4, the value of RF increased with increasing the temperature. It was clear that both values of '2co+n2 and 2co+H2 also increased and the rise of Tg depended on the increase of '2x2 than 'co. This meant that the rate of hydrogen reduction was apparently accelerated with temperature increase. 4. Effect of Oxidizing Gas and Methane Concentration

In the viewpoint of the evaluation of process efficiency including the production of reducing gas and the recycle of exhaust gas, optimum reducing

gas composition must be determined in the industrial process for manufacturing the reduced iron. Runs 4 to 13 were attempted to examine the effect of con-centration of CH4 and of oxidizing gas in the reducing

gas on the operational results.

Table 8. Heat

side

of reaction of

reactions.

stepwise reduction and

Fig. 4, Relations of final utilization degree

(Ti).8)

reduction degree (RF) and gas

(r2) with blowing gas temperature

Research Article

Transactions 'SIT, Vol. 26, 1986 (771)

By plotting RF and v with the value of R V defined in Eq. (26), Fig. 5 was obtained. A relationship was found in the figure for the different operational conditions:

RV = (Yco+Yf12)/(Yco2+YH2o) ............(26)

In Fig. 5, RF decreased sharply in the region of R V less than 20, but both values of RF and ry were almost constant in the region of RV over 20. It may be considered that the relation between RF and R V depended on the intrinsic reducibility of the pellets used. Consequently, both R1, and R V had about 0.87 and 20, respectively, in the case of N pellet used. On the other hand, ~7co+x2 was almost equal to

in the R V range over 20 as shown in Fig. 5. This expressed that reducing gas was effectively used by the reduction of the pellets under the conditions. Therefore, R V of 20 indicated optimum operational condition in the case of the pellet used considering the constant RI, obtained and the recycle of exhaust

gas in an industrial process.

2. Reduction Fashion o f the Pellets It is estimated that the reduction of the pellets

decending in the actual shaft furnace having distribu-tions of temperature and of gaseous concentration shows intermediate fashion between idealized fashion having a reaction interface observed at the high reduction potential and the one having a reaction zone such as step-wise reduction at the low reduction

potential. The reduction model of a single pellet which is included in the mathematical model for the shaft furnace, should be selected according to the observation of reduction fashion.

From this standpoint, the reduction fashion of the

pellets was observed. Photograph 1 shows the cross sections of the pellets sampled from the shaft furnace operated at ordinary pressure (Run 3) and at high

pressure (Run 4). Although the interfaces between Fe2O3 and Fe3O4 and between FexO and Fe were observed clearly in the pellets, the interface between Fe3O4 and FexO was unclear. However, thin layer of FexO was observed by etching the sample between Fe3O4 and Fe layers as previously reported.' This fashion mainly caused by the low reduction rate at the reduction step from Fe3O4 to FexO under 973 K which was observed by the authors.11~ It was found in Photo. 1 that the reduction of the pellet at ordi-nary pressure proceeded with multi-interfaces. This

proved the application of the multi-interface reaction model to the reduction of pellets in the shaft furnace.

Fig. 5. Variation of final

utilization degree

reduction degree

(r;) with RV value.(R1) and gas

Photo. 1. Cross sections at each level bed for Run 3 4 (b).

of the pellets in the moving

(a) and Run

Research Article

( 772 ) Transactions ISIJ, Vol. 26, 1986

On the other hand, in the high pressure reduction for Run 4, the pellet of No. 7 in Photo. 1 (L=105 cm) seems that the reduction has finished apparently in spite of the reduction degree of 0.722 which was also found in the distribution of reduction degree shown in Fig. 3.

Photograph 2 shows the microscopic view of the

pellet sampled from Run 4. FexO was found to remain almost uniformly within the grain constituting the pellet. Because the reduction at high pressure

proceeded somewhat homogeneously, another reduc-tion model except the multi-interface reaction should be adopted to the mathematical model for a shaft furnace.

3. Crushing Strength of the Pellets in the Shaft Furnace

In the hydrogen reduction,5> the authors reported that the crushing strength of the pellets in the shaft furnace decreased with the progress of the reduction from Fe2O3 to Fe3O4. In the reduction with gas mixture in the present work, advanced informations were obtained as follows :

For an example, Fig. 6 shows the distribution of crushing strength (Q],) and of reduction degree (R~,) of the pellets sampled from Runs 9 and 14. It was found from Fig. 6 that the strength of the pellets were considerably different with each other, especially, the decrease of the strength was slight in case of the rapid reduction from Fe2O3 to Fe3O4. By inspecting the retention time in the range of reduction degree from 0.11 to 0.33 (tR=o.11~o.33) and average crushing strength (op) of the pellets when it comes to the lower part of the shaft furnace, the relationship was obtained between them from various experiments as shown in Fig. 7. The strength of the pellets was not found to indicate marked decrease if the retention time was under 3.0 X 103 s.

These phenomena were caused by the following reasons. The Fe layer which was formed at the

periphery of the pellet under high reduction potential prevented a decrease of the crushing strength by

surpressing the volume expansion associated with the reduction step from Fe2O3 to Fe3O4. In the case of

low reduction potential, the strengh of the pellets

which was lessened in the upper part of the shaft furnace, did not recover by the formation of Fe layer in the lower part of the furnace because of the different

fashion of the reduction. It was hitherto explained that the strength of the

pellets in the shaft furnace depended on the physical

properties of unreacted pellets; however, Fig. 7 shows that it was controlled by the reducing conditions.

Therefore, the decrease of the strength and degrada-tion of the pellets will be improved by the selection

Photo. 2. Micro-photograph of the pe

ent levels for Run 4.

llets sampled from differ-

Fig. 6. Longitudinal distribution of crushing strength (Qr) and reduction degree (R) for the pellets obtained by

Run 9 and Run 14.

Fig. 7. Relation between average values of crushing strength (6r) and retention time of the pellets in the range of reduction degree from 11 to 33 %.

Research Article

Transactions ISIJ, Vol. 26, 1986 (773)

of optimum operation conditions.

VII. Computed Results

A simulation computation estimating the distribu-tion of process variables was carried out by using the mathematical model for the shaft furnace mentioned

previously. Computed results were compared with the experimental results to investigate the soundness of the fundamental equations and the numerical computation technique in the mathematical model. Six modes of computation for different rate param-eters were attempted to examine the appropriateness of the reaction rates as shown in Table 9. The modes 1 and 2 in Table 9 represent the cases of the different rate parameters for the reduction of pellets and of those for the side reactions respectively.

The reasonable simulation of the shaft furnace was obtained not only by mode 1-(3) but also by mode 2-(3) in atmospheric pressure operation, because methanation reaction did not proceed in the case. One of the simulation result was shown in Fig. 8. Simulation results in other modes were shown in the previous paper18j and not reproduced here. Con-sequently, the reduction behavior of the shaft furnace at ordinary pressure can be simulated well by the mathematical model considering only kinetic of the water gas shift reaction or both kinetics as a side reaction. Mathematical simulation for the high pressure reduction by the shaft furnace will be conducted by the authors in near future.

VIII, Concluding Remarks In order to clarify the effect of the operational

conditions on the operational results of the shaft furnace, some series of reduction experiments of iron oxide pellets by gas mixture were carried out by using the pressurized shaft furnace under the differ-ent operational conditions. On the other hand, a one dimentional mathematical model which con-tained the kinetics on the reduction of iron oxide by gas mixture and side reactions, was developed and used for the simulation of the shaft furnace.

In summary, the important results obtained by the experiments are as follows :

(1) In the reduction by gas mixture, higher temperature distribution and higher final reduction degree (RF) were observed in comparison with hydrogen reduction because of the exothermic behav-ior of the water gas shift reaction.

(2) In the high pressure reduction, RF of the pellets was decreased by the lower reduction potential of reducing gas due to the formation of CH4, H2O and CO2.

(3) R~, and gas utilization (r~) were increased with increase of inlet gas temperature.

(4) In the relations of RF and v with RV, RF was sharply reduced in R V value less than 20, but both values of RF and were almost constant over RV value of 20. Therefore, R V value of 20 indicated optimum operational condition in the case of the

pellet used. (5) Reduction of the pellets proceeded topo-

chemically at ordinary pressure but rather homogene-ously at high pressure.

(6) The strength of the pellets did not decrease remarkably if the retention time was short in the range of reduction degree from 11 to 33 %.

Furthermore, the important results obtained by the simulation are as follows :

(1) The mathematical model considering the reduction of pellets and side reactions gave the reasonable simulation results for the shaft furnace at ordinary pressure.

(2) The evaluation of rate parameters for the

Table 9. Calculation modes

matical model for

applied to the

shaft furnace.

mathe-

Fig. 8. Comparison between observed data and calculated curves by using the mode l-(3) for Run 11.

Research Article

( 774 ) Transactions ISIJ, Vol. 26, 1986

reduction of iron oxide pellets was important factor for the simulation of the reduction behavior in the shaft furnace, especially in the case of the retardation in the reduction step from Fe304 to FexO under 973 K and in the range of high reduction degree.

(3) Side reactions should be included in the mathematical model to obtain the good simulation results for the reduction with gas mixture in the shaft furnace.

(4) The equilibrium of water gas shift reaction was not reasonable assumption.

Nomenclature

cy, c~, : Molar specific heat of gas and solid, respectively (cal/mol • K)

D~, DG : Effective diffusivity and molar diffusion coefficient, respectively (cm2/s)

D1 : Inner diameter of shaft furnace (cm) doi : Removable oxygen concentration within the pellet in the each reduction step

(mol(0)/g(pellet)) d~ : Diameter of pellet (cm)

G : Molar flow rate or flow rate of gas

(mobs) or (m3(STP)/s) G; : Molar flow rate of a component gas

(mobs) GM : Mass velocity of gas (g/cm2 • s)

g~ : Force conversion factor 980.665 (g • cmf g•s2)

-4H: Heat of reaction (cal/mol) h : Heat transfer coefficient between solid

and fluid (cal/cm2 • s • K) kM, kw : Equilibrium constants for formation and

decomposition reaction of CH4 and for water gas shift reaction respectively,

KM (atm-2), K(-) kci : Reaction rate constant of each reduc-

tion step (cm/s) kM, k~r : Reaction rate constants of formation

and decomposition reaction of CH4. kM (molf s • cm3 • atm4), k~s (mol/s • cm3 •

atm2) kw, kw : Reaction rate constants of forward and

reverse water gas shift reactions (mobs cm3 • atm2)

L: Effective height of shaft furnace (cm) 0 x : Fraction of removable oxygen within the pellet (-)

P: Total pressure (atm) or (MPa)

p : Partial pressure (atm) R : Gas constant 8.3146 x 10-3 (kJ/K • mol)

R1, R7,, Ri : Reduction degree of final, overall and each reduction steps (i= 1: Fe203 - *

Fe304, i=2: Fe304 --> FexO, i=3: Fey 0 -p Fe, i=4: Fe304 -> Fe)

Rer, : Reynolds number Ti : Rate of formation of gas (mobs)

S : Cross area of shaft furnace (cm2 ) Ta, TTg, TTS : Temperatures : ambient, gas and solid,

respectively (K) U: Overall heat transfer coefficient (calf

U:

vi :

VP:

W:

WK :

vi:

z:

Ul, :

~:

vi:

cm2•s•K) Linear velocity (cm/s) Reduction rate (molf s) Volume of the pellet in control volume (cm3 ) Flow rate of pellet (g • mobs) or (kg (Fe)/s) Flow rate of each component of pellet

(g•mol/s) Molar fraction of each component gas Distance from the top of the furnace

(cm) Average value of crushing strength of the pellet (kg) Voidage of the bed (-) Gas utilization degree (%)

Acknowledgements Our previous paper published in Ironmaking Pro-ceedings, 43 (1984), p. 485 has some wrong figures in Tables 3, 5, 7 and 8. They are corrected in the corresponding Tables of 3, 5, 6 and 7 in the order in this paper.

REFERENCES

1) Y. Hara, M. Sakawa and S. Kondo : Tetsu-to-Hagane, 62

(1976), 315. 2) T. Murayama, Y. Ono and Y. Kawai : Yetsu-to-Hagane, 64

(1978), 1518. 3) T. Yanagiya, J. Yagi and Y. Omori: Ironmaking & Steel-

making, 6 (1979), 93. 4) Y. Takahashi and R. Takahashi: The 8th Japan-USSR

Joint Symposium on Physical Chemistry of Metallurgical Processes, ISIJ, Tokyo, (1981), 215.

5) R. Takahashi, Y. Koyabu, M. Ishii, M. Ishigaki and Y. Takahashi : Tetsu-to-Hagane, 66 (1980), 985.

6) Y. Hara, T. Wakabayashi and N. Otsuki: The 8th Japan USSR Joint Symposium on Physical Chemistry of Metal- lurgical Processes, ISIJ, Tokyo, (1981), 229.

7) R. Quintero and E. Martinez V.: Ironmaking Proceedings, 43 (1984), 501.

8) R. Takahashi, M. Ishigaki and Y. Takahashi: Yetsu-to- Hagane, 69 (1983), S771; Trans. ISIJ, 24 (1984), B73.

9) R. Takahashi and Y. Takahashi: Tetsu-to-Hagane, 69

(1983), 5773: Trans. ISIJ, 24 (1984), B75. 10) Y. Hara, M. Tsuchiya and S. Kondo: Tetsu-to-Hagane, 60

(1974), 1261. 11) M. Ishigaki, R. Takahashi and Y. Takahashi : Bull. Res.

Inst. Min. Dress. Met. Tohoku Univ., 36 (1980), 23. 12) Y. Omori, J. Yagi and T. Yanagiya: The 7th Japan-USSR

Joint Symposium on Physical Chemistry of Metallurgical Processes, ISIJ, Tokyo, (1979), 20.

13) Y. Hara, M. Sakawa and S. Kondo: Tetsu-to-Hagane, 62

(1976), 324. 14) M. Ishigaki, R. Takahashi and Y. Takahashi : Bull. Res.

Inst. Min. Dress. Met., Tohoku Univ., 38 (1982), 35. 15) 0. Kubaschewski, E. L. Evans and C. B. Alcock: Metal-

lurgical Thermochemistry, 4th Ed., Pergamon Press, (1967). 16) A. Rist and N. Meysson: J. Metals, (1967), April, 50. 17) K. Narita, D. Kaneko, Y. Kimura, Y. Takenaka, H. Tanaka

and Y. Inada : Tetsu-to-Hagane, 67 (1981), 508. 18) R. Takahashi, Y. Takahashi, j. Yagi and Y. Omori:

Ironmaking Proceedings, 43 (1984), 485.

Research Article