ISIJ International, Vol. 39 (1999), No. 7, pp. 705-714

Operation

Bed Typeand Design of Scrap Melting Process of the Packed

Takaiku YAMAMOTO.Yutaka UJISAWA.Hiroaki ISHIDAand Kouji TAKATANICorporate Research & Development Laboratories, Sumitomo Metal Industries. Ltd., 16-1 Sunayama

Kashima-gun, Ibaraki-ken, 31 4-02 Japan.

(Received on February 16. 1999.• accepted in final form on April 14. 1999)

Hasaki-machi,

Acoke Packedbed type Scrap Melting process (PSM)with simultaneous high-rate injection of oxygenand pulverized coal was proposed which aims to achieve high heat efficiency, high productivity, high Feyield and low refractory erosion with using not only shaft type furnaces but also basic oxygen converter(BOF) vessel,

The experiments were madeto verify the usefulness of the proposed process using the modifiedexperimental blast furnace of I ton pig per tap scale and the modified experimental BOFof IOton pig pertap scale. A 3-dimensional mathematical simulation model was also deve[oped in order to elucidate thephysical and chemical conditions inside the furnace on the basis of transport phenomenaand to makeabasic design of commercial scale plant. Obtained results were summarizedas follows:

1) 100'/. steel scrap can be melted into well carburized and well desuipherized pig iron using both I tonscale blast furnace and IOton scale BOF.

2) The heat efficiency of about 80'/•, the productivity per unit furnace volume of about 50t/d • m3, Fe yieldof morethan 99•/• and the erosion rate of refractory of less than 0.5 mm/chargewereachieved throughoutthe experiments.

3) These results were well justified by the 3-dimensional mathematlcal simulation model,4) Abasic design of commercial scale plant wasalso performed by the use of the model.

KEYWORDS:scrap melting; coke packed bed; pulverized coal; oxygen; post combustion; experimentaiblast furnace; experimental basic oxygenconverter; mathematical model; scale up; commercial scale plant.

1. Introduction

Recycling of iron scrap is expected as acountermeasurefor the effective utilization of resources and against the

global warning by carbon dioxide (C02)'1) In recent

years the amountof old scrap is increasing steadily byabout one million tons per year in Japan. This tendencywould continue in the future.2) Therefore the integratedsteel plant should develop the scrap melting processimmediately with use of its advantages. Twomethodswhlch use the blast furnace3) and basic oxygenconverter(BOF)4) are consldered as the method to utilize the

equipments in the plant. Although the equipmentof BOFis simpler than that of blast furnace

,there is someroom

to improve the heat efficiency and refractory life.

Therefore, we should choose the best methodafter the

various investlgations has been performed in the future.

Fromthese points, the newscrap melting process (PSM:Packed bed type Scrap Melting process)5) which canapplled the equipment of BOFwas developed bySumitomoMetal Industries,Ltd.

As the melting furnace is a kind of vertical furnacewhich has coke packed bed6) and its heat exchangebetween gas and solid is countercurrent flow, it is con-

sidered to be superior to in-bath smeltlng with use of

BOFin regard to heat efficlency and refractory life.

In this paper, we describe a summaryof the scrapmelting tests madeto verify the usefulness of the proposedprocess using a modified experimental blast furnace of

i ton pig per tap scale and a modified experimental BOFof 10 ton pig per tap scale first, Next, we discuss a3-dimensional mathematical simulation modelwhich wasalso developed in order to understand the internal scrapmelting behavior inside the furnace and establlsh thetechno]ogy of scaling up the process.

2. Concept of PSMProcess

Figure I shows the concept of PSMprocess sche-matically. The melting furnace has a charglng systemsof scrap and coke and discharging exhaust gas at the

top of the furnace with plural primary tuyeres in its

10wer portion and plural secondary tuyeres in the side

wall at a level above the primary tuyeres. The furnaceis equipped with a tapping hole for hot metal and slag

at a levei between the primary tuyeres and bottomtuyeres. The air or hlghly oxygen-containing gas is

b]own into the furnace of high temperature coke packed

705 1999 ISIJ

ISIJ International, Vol. 39 (1 999), No. 7

Lowoff-gas temperature

O Coke HtirlJ~(Hlgh heat etflaency)

II Scrap

Packedbed(High iron yield)

secondary tuyeres0= -~ c0+1120==co,

o,,coal -- """"'coke>'.'.'.'.'.primary tuyeresc+1120, =co(High reducing ability)

center bottom tapping 0~ (cao)Bottom tuyeres

Fig. l. concepl orule PSMprocess.

bed from the prlmary tuyeres if necessary with pulverizedcoal Injection and then coke and pulverized coal are burntin front of the tuyeres according to Eq. (1).

C+ll202=C0+294Mcal/kmol C .........(1)

The scrap is melted by the sensib]e heat of the high

temperature reducing gas produced by the combustion.

On the other hand, the scrap is preheated by the gasstreaming upward after melting the scrap. The postcombustion expressed by Eq. (2) is caused by the air or

oxygen- contalning gas injectlon into the shaft from the

secondary tuyeres promotes the scrap heating.

CO+ l/2 02=C02+67.6 Mcallkmol-CO ......(2)

Features of the PSMprocess are summarizedas follows.

l) As countercurrent heat transfer occurs in the furnacewith the packed bed of scrap and coke, the total heatefficiency(over 80 o/o: see Fig. 12) remains higher thanthat (about 70-75 o/07)) of EAF(Electric Arc Furnace)

or that (about 60 ol08)) of in-bath sme]ting method.2) The iron yield is higher than that of in-bath smelting

methodby the packed bed method.3) As the refractory is protected from high temperature

atmosphere in the furnace by the packed scrap andcoke, its refractory erosion is lower than that ofin-bath smelting method.

4) Blast furnace coke with a large quantity ofpulverizedcoal instead of high quality as foundry coke which is

necessary for conventional cupo]a is available for

PSM.5) Although the high-grade coke with low reactivity and

large size is burnt nearly to C02(C+02=C02)inconventional cupola, the ordinary coke for blast

furnace is burnt to COaccording to Eq. (1) in PSM.Therefore, although reducing ability in the lower partof the furnace in conventional cupola is low, it is sohigh in PSMthat the hot metal whosequality is equal

to that of blast furnace is produced with highcarburization and high desulfurization using 1000/0

stee] scrap.6) The fuel efficiency and productivity is increased by

the post combustion according to Eq. (2) with the air

or oxygen-containing gas injection from the second-

ary tuyeres muchmore than these of blast furnacemethod.9)

Modified experimental blast furnace Modified experimental BOF

Heat size(ton pig / tap) 1 Heat size(ton pig / tap) 10

Inner volume (m3) 2.2 Inner volume (m3) 6.9

Oxygensupply(Nm3/hr) 350 Oxygensupply(Nm3/hr) 1800

7 o 13 oc

,t

Secondarytuyeres ,:,

~ 1840

Secondary{tuyeres ':,~o 8

Primarytuyeres 1e40

Primary ~p

tuyeres ~14

Taphole Taphoie Bottomgooc tuyeres

Fig. 2. Configul Itlon of the cxpelmient ll fulnaccs

Table l. Chemical analysis of raw materials.

c H o N S AshCoke 87 0.2 1.1 11 10.6

Coal 75 4.5 11.o7.7 1.4 o6

7) Although a shaft type furnace such as conventionalcupola is suitable for the furnace body, a conventional

BOFconverter is also avallab]e for its body.

3. Experimental Method

3.1. Experimental ApparatusAfter the fundamental functions of PSMprocess was

proved using a modlfied experimental blast furnace of

l ton pig per tap scale, the scaling up test wasperform-ed using a modified experimental BOFof 10 ton pig

per tap scale. Figure 2 schematically illustrates theconfiguration of experimental furnaces. The I ton fur-

nace has three primary tuyeres installed at 120 degreeclrcumferential intervals and three lower and uppersecondary tuyeres were instal]ed above the primarytuyeres, respectively. Onthe other hand, the 10 ton BOFhas four primary tuyeres installed at 90 degree cir-

cumferential intervals and six lower and upper second-

ary tuyeres were installed above the primary tuyeres,respectively. Furthermore, two bottom tuyeres wereequipped in order to control the temperature and Si

content of hot metal.

3.2. Material Condition

Tab]e I shows the properties of raw material used in

the present experiments. Sized coke for blast furnace andnon-coking coal with volatile matter of 350/0 after

pulverized into the size of 800/0 under 200 meshwereused as fuels. Limestone or calcium oxide and serpentine

were used to adjust the basicity and MgOcontent ofslag, respectively. Steel scrap whosepurity of iron has99 masso/oFe wasused in the tests.

3.3. Operational Condition

Table 2showsthe standard operational conditlons for

the I ton furnace and the 10 ton BOF.

O1999 ISIJ 706

ISIJ International, Vo[, 39 (1999). No. 7

Table 2. Experimental conditions.

hour/campaignOeration timeto n/tapTa in

Avera e ta to ta hour/tap-tap

Scarp Weight kg/pieoeBuik density kg/m

Coke Size mmBulk density kg/m

Total oxygen Nm3/hbowinCoke/Fuel Ratio

Production tHM/hrate

2410

imaxIoo3.0 x Ia3

50 - 80500

max1800

1 - 0.25

approxi5

Table 3. Chernical analysis ofhot luetal and slag.

~~

>oC:

O.5~E(1)

cVQ)

I

1oo

80

60

40

CSi

MnPS

Temperature

(CaOJSi02)

Volume

ee

Oft gas temp

500 'C

1500 'c

e : Actual data of PSM

- : Theore~cal line

20025 50 75 1OO

Post COmbustionratio (%)

Fig. 4. Influence of post combustion on heat emciency.

Heat in ut iOO%

Coke Others

c:o

c~E~~ 500

!:coc~;200OOE

z~- 100a'~ OO

400

+'(D(a*~

- 200~5~,:'!~LL~

o

-e-..~,, o oo~i•--~0e-~_.~~

e 1ton furnace

O 10ton furnace\~e ee_o ee:~:~~c~:o(~o

.__e-t~!cr

30 50O IO 4020Post combustion ratio('/•)

Fig. 3. Influence of post combustion on fuel rate and oxygenconsumption.

4. Experimenta] Results

The tests on both of the furnaces were performedentirely smoothly. As a results It was confirmed that it

is sufficlently possible to apply PSMprocess to BOFtypefurnace. As the lower part of the furnace wasmaintainedat high temperature and reducing atmosphereby the cokepacked bed, the hot metal whosequality is equai to that

of blast furnace was produced with high carburization,

high desulfurization and trace Si content. The averagecompositions of pig iron are shownin Tab]e 3. The fuel

and oxygen consumption in the tests were summarizedin Fig. 3. It was verified that the fuel and oxygenconsumption was decreased by the increase of postcombustion. Figure 4showsthe relationship betweenpostcombustion and heat efficiency in the 10 ton furnace.

The heat efficiency of more than 800/0 Wasachievable

whenthe post combustion ratio reached nearly 30 o/o andthe off gas temperature was less than 500'C becausethe heat generated by the post cornbustion was oper-ated effectively to heat the scrap. Figure 5showsan ex-arnple of heat balance In the 10 ton furnace when the

post combustion ratio wasequal to 300/0. At that time,

scrap of 7.5 ton per charge wasmelted comp]etely within

53olo

Sensitive andlatent heat of

Hot meta

Fig. 5.

Sensitive heatof siag

34'/. I '/.

12'

Sensitive heat ofoff gas andheat loss

Chemical heat of off gas

Exampleof heat balance.

around 30 minutes and the production rate of molteniron was equlvalent to around 50t/d•m3. The ironyield of more than 99 o/o Wasobtained and the averagerefractory erosion was maintained at low erosion level

of less than 0.5mmper charge through the all test

period in both of the furnaces.

C02+H20Post combustion (o/o)= C0+C02~H20X 100 (3)

Hest efficiency (o/o) =Heat of hot metal

Heat inpu']Chemical heat of off ~as X 100 (4)

5. Development of a 3-Dimensional MathematicalSimulation Model

PSMprocess is a packedbed type reactor such as blast

furnace or cupola which Is a continuous countercurrentprocess between gas and solid or liquid with simultane-

ous heat and mass transfer as wel] as manychemicalreactions. Therefore, a lot of attempts, Iaboratory scale

experiments and dissection investigations of the reactorshave been studied for the better understanding of thesecomplicated phenomenaconcerning physical and chemi-cal mechanisms

. Onthe other hand, by the use of these

works, all klnds of mathematical models have also beenconstructed in order to estimate quantitatively the

phenomenaoccurring in the reactor. In recent yearscomplex analysis has becomepossible because of the

improvement of the ability of the digital computer andnumerical anaiysis technology. A 3-dimensional mathe-

707 O1999 ISIJ

ISIJ Internationai, Vol,

matical simulation modello.11) that can describe threedimensional and unsteady state of the internal behavi-

or that could not treated by the previous mode]sl2 16)

wasdeveloped in order to design the process and improveits operation systematically based on understanding dy-namical and physical transport phenomenawith chemi-cal reactlons inside the reactor.

5. I .Modeling

The governing equations to describe the internalbehavior was composedof three dimenslonal and dy-namic equations in order to design the reactor profile,

that is, to study the effects of arrangement of tuyeresand deflection of gas and solid fiow in the circumferentialand radial directlon and radial direction on the op-erationai results and to design the unsteady state oper-ational method in the sameway as the actual opera-tlon. The following assumptions and simplification weremadein the model.

1) Three phases (gas, solid and liquid) exlst inside thefurnace.

2) Species of gas are considered CO. C02, H2. H20,N2and 02'

3) Species of scrap and coke are Fe, etc and C, etc.respectively.

4) Specles of liquid are Fe. Cand etc.

5) Considering main chemical reactions occurring in

the reactor wlth reaction rate equ'ations.6) Ergun equatlonl7) js applied to gas flow in the

packed bed.7) Kinematic modell8) is applied to solid flow of

burden materials.8) Liquid is assumedto moveto the gravity direction

wlth a constant dripping velocity.

9) Density of gas is calculated according to ideal gasequation.

lO) Temperature of scrap and coke of so]id phase is

assumedto be same and that of metal and slagin 1lquld phase Is assumedto be same.

1l) Heat ba]ance is considered for each phaseby takingaccount of the heat reactions, heat exchange, heatconduction and heat loss. However,heat conductionis not considered for liquid phasein dispersive phase.

5.2. Governing Equations

The governing equations are written for mass, mo-mentumand energy in the following form.

5.2. I .MassBalance

Dr8, n .; ..1

_~ Jr J_~L! ~ (5)Dt = aj div(pjDjgrad ~ji) +Mi Rate,*aji,,...

,'

Mass balance equation is written as follows by sum-mation of massfraction at each phase :

ND(8jpj) ~ ~ ••••••(6)= (Mi Rate,,)aji,*..........Dt i = I ,*

Where, it is necessary to conslder mixing and diffusion

phenomenaamonggases in order to eva]uate the effect

of post combustion reaction inside the packed bed onthe reactor efficiency appropriately whena relatively lowheight reactor such as BOFtype furnace is used for the

Q1999 ISIJ 708

39 (1999), No. 7

AC

EEo~~

Air

=~N2=~

300 mm

Measuring 02 Concentrationat EachPoint

Saction pipe for gas sampling

Air

=96 Nms/hSecondarytuyere

N=~-= 96 Nm3/h

Primary tuyere

Fig. 6. Experimental apparatus for the investigation of the

mechanismof gas diffusion through the packed bed.

~~

a~

oa:

loe)

.~~

E~oo(oa:

80

60

40

(Measuredratio of mixed area)..--- (dp= 22,5)

•' ,."---(dp= 17.5)Calculated line

d= 22~d= 17,5

d* 4,0

200

Fig. 7.

4 8 12 16 20Pepnumber(= dp ' Ug/Dj )(-)

Effcct of partlcle diameter on the gas diffusion.

PSMprocess. The diffusion of gas phase is consideredin the first term of right hand side of Eq. (5). The onlydimensionless numberto govern the gas diffusion is Pecletnumber(Pep) that meansthe ratlo of bulk masstransportand masstransfer by the diffusion as shownin Eq. (7).

Pe=c!p• Ug

..........(7)P Dj " "

Pep in packed bed is known to be a constant value.19)

However, the Pep in three dimenslonal state has not yetbeen investigated. Therefore, a cold model test wascarried out in order to Investigate the mechanismof gasdiffusion through the packed bed. The experimentalapparatus is shown in Fig. 6. After packing particles,nitrogen and air were injected from the primary andsecondary tuyeres, respectively and the oxygen con-centration was measured at each point In order toevaluate the mixing and diffusion phenomenaamonggases quantitatively. The degree of the mixing and dif-

fusion amonggases (called mixed area) was evaluatedby the area of oxygen concentration range from 5to 15

volumepercent. ThePepWasdetermined to comparetheexperimental mixed area with the calculated one on thebasis of the relationship between mixed area and Pepsimulated by the model as shownin Fig. 7. It is apparentfrom Fig. 8that the calculated values by the determinedPep (=4) agreed well with the experimental ones.20)

5.2.2. MomentumBalance(1) GasPhase

a(8gPgUg)

= _ eg grad PL 8g(.fl +,f2 18gPgUgl)(8gPgUg)at

.(8)

ISIJ International, Vol, 39 (1 999), No. 7

Calculated

A O A A O A A Odp(mm)= 4(3-5) 12.5(lO-15) 22.5 (20-25)

Fig. 8. Determination of Peclet numberin the packed bed.

A O A A O A A o A

20 20

20 20 20 20J5 ~)

15

5 5 5 55

Peps54.0 Pep=4.0 Pep=4.0

lculatedc Measured Calculcted c Measured Calcu lated c Measured

5 5 5 5 5520 f5

2

*,fl= 150:,ip~:)'

f2=1. 75l-eg) l8gipcll] e~pg

A

(2) Solid PhaseThe kinematic model which has been confirmed to be

useful through a 3-dimensional cold modei experimentfor burden descent21) was applied as the equation ofmotion for the solid. The governing equation of ki-

nematic model is given by the following equation whent~e coordinate of Xj is in the gravity direction:

U,i =BaU~j aU~j

aXi ' U~k=B ..........(9)

aXk ""

(3) Liquid PhaseMotion of liquid was assumedto drlp vertically as a

plug flow as follows:

U .=const., Ul* Ulk O ..........(lO)IJ

52.3. Energy Balance

Dr8 n C .T.1\ JrJ pJ J) =8jdiv(kjgradTj)

Dt

+~{a. u. (T,,,-T.)}+RH. ...(11)J''' J'** J J

Where,

D( ) a( ) a( ). RHj= ~Rate,,( - AH)n

.---- =--- - - + Ui*' J''Dt at aX ,,

5.3. Chemical Reactions

Main reactions occurring in the furnace were con-sidered as reaction rate equations in the model. Thereactions are shownin Table 4. It is possible to evaluatethe effect of the post combustion in the packed bed thatis one of the most important factors to govern PSMprocess with gasgasreactions No. 2, 3by oxygeninjected

from the secondary tuyeres and solidgas reaction No. 1as shownin Table 4. Temperature recovery method22)

wasapplied to the melting of scrap (reaction No. 7).

5.4. Process ParametersProcess parameters that are necessary for the sim-

ulation were applied to the reference values shown in

Table 4. It should be noted that liquid-solid heat transfercoefficient at the dripping zone had to be reduced to halfof the original value25) quoted in the model based on

Table 4. Chemical reactions and process parameters in the

model.

n Chemical reactions Rate~ Ref.

1 C(s) + 1/20=(g) - CO(g) 23

2 CO(g) + 1/20,(g) - CO,(g) Mixing control

3 H2(g) + 1/202(g) - H20(g) Mixing control

4 C(s) + C02(g) - 2CO(g) 24

5 C(s) + H20(g) - CO(g) + H2(g) 24

6 H20 + CO = H2 + C02 E uilibrium

7 Fe(s) - Fe(1) (Melting rate) 22

8 3Fe(i) + C(s) - Fe3C(l Equilibrium

Process arameters Smbol Ref.Effective thermal conductivityfor solid hase

ks 27

Thermal diffusion coefficientfor as hase

kg 28

Gas*solid heat transfer ooetficient Ug 25Gas-li uid heat transfer coefticient Ugl =0Li uid-solid heattransfer coefficient us 26Gas*solid effective contact area as* = aall-al

Gas-li uid effective contact area asl 29Li uid-sold effective contact area aJS 30Where, aal : Total contact area of solid phase

start

Set grids of the fumace

Set init'al oondition

Solve flow fields for gas, solid and liquid phase

Solve massconservation equations for gas, solid andLiquid phase

Solve energy conservation equations for gas, solid andLiquid phase

Set newgas, solid and liquid state variables

Aftertime 'nterva llt

End

Fig. 9. Flow chart of the solution procedure.

analysis of operational results.

5.5. Boundary Condition1) Scrap and coke were assumedto be continuously

charged into the furnace from the top with mixedcharging or separated charging and the liquid wasassumed to be continuously discharged from the

bottom of the furnace.2) Blowing conditions are set at the primary and sec-

ondary tuyeres in the sameway as the actual oper-ation.

3) Heat transfer into the refractory is described by onedimensional heat conduction model in the thicknessdirection.

5.6. Solution Procedure

Thedifferential equations to be solved numerically areabove mentioned Eqs. (1)-(6) and these equations aredynamically calculated by the time difference At on thebasis of the appropriate boundary condition and theinitial condition. Thefiow chart of the so]ution procedureIn this model is shownin Fig. 9. As computation cells,

709 @1999 ISIJ

ISIJ International, Vol.

staggered grids were used and a finite difference methodbased on Gauss's divergence theorem31) wasapplied to

represent the reactor shape exactly. The analysis algo-

rithm of gas flow was applied SOLAalgorithm32) thatis frequently used for numerical fluid dynamics. Massand energy conservation equations for each phase werediscretized by the implicit schemeandsolved by the point

Table 5. Operatronal conditrons of I ton furnace

Items Unit ConditionsBurdenScarp size mm i 50~] x 20tCokesize mm 20-50Charging pattern hlb(edStook level m 1.6

BiowinPrimary tuyeres pieces 3p02 Supply Nm3/hr 207N2SUpply Nm3lhr 61Coal/02 1.O

Boshgas volume Nm3/hr 622Flametemperature ~c 2847Secondarytuyeres pieces 3pAir supply Nm3/hr 207Productivit

Produchonrate tHM/day/m3Nm3/hr/m3

i802 supply 114

Secondarytu yere

Primarytu yere

Fig. lO. Gl'ids of I toll furnace.

39 (1999). No. 7

SORmethod.33)

6. ComputedResults

Thevalidity of the modelwasevaluated by comparisonbetween the operational results of the above mentionedtest furnaces and the computedresults and the internal

behavior of the process wasdiscussed. As a results, theestimation of the internal behavior of PSMand theoperational design of commercial scale PSMprocessbecamepossible by the model.

6.1. Results on the I Ton Test Fun]aceIn order to study the verification of the present

modelas a total simulator, the computer simulation wasperformed for a practical operation of the I ton test

furnace and the calculated results were comparedwiththe actual test results. Theoperational condition for thecalculation is shownIn Table 5. Thecomputational grids

are shownin Fig. lO. The internal behavior of the fur-

nace calculated by the model is shownin Fig. 11. Thecalculated gas flow pattern in the furnace showsalmostplug fiow in the region above the primary tuyere level

except for the zone in front of the tuyeres. Thecalculatedsolid flow was also found to be plug flow in the regionabove the primary tuyere level. The solid velocity nearthe primary tuyere is high because the volumechangeofsolid phase is occurred as a result of the consumptionof coke by the carbon monoxide combustion and themelting of scrap (the melting point was assumedto be

1400'C). Onthe other hand, the velocity at the central

zone of the furnace near the primary tuyere level is lowin comparlson with that in the peripheral zone. The heatexchangebetween gas and solid is performed promptlyat the upper part of the furnace and the solid heatingis promoted by the post combustion of COgas causedby the air Injection from the secondary tuyere as shownin C02profile of Fig. I I .

Theflow of liquid formed at theme]ting polnt has the distribution which is proportion-al to the solid velocity. The temperature of the dripping

C 1999 ISIJ

500

Gasfiow and Solid flow and Liquid flow and CO C02temperature(~C) temperature(~C) temperature(~C) Gasconcentration(vol.'/.)

Fig. 11. Internal behavior of I ton furnace c'alculated by the present model.

71 o

02

ISIJ International, Voi. 39 (1999), No. 7

1500

oa)

~ 1000~o

a)C~Ea)

H 500

'\ OA : Measured\

- : Calculated\ \ - : Calculated\ \ \ ~\ \

\\\lGas temperature

\ \\~\

\~ \\ \ \

Solid temperature \

O o 0.4 0.8 1,2 1.6

Height from primary tuyere (m)

Fig. 12. Comparison of measured longitudinal temperaturedistribution with calculated one in the I ton furnace.

Table 7. Operational conditions of 10ton furnace.

Items Unrt ConditionsBurdenScrap size mm 217[] x 80tCokes'ze mm 40-80Charging patternStock evel

separatedm i .9

BlowinPrimary tuyeres pieoes 4p02 supply Nm3/hr 779N2supply Nm3/hr 217Coal/02 0.65LPGsupply Nm3lhr 37.5Boshgas volume Nm3lhr 226iFlametemperature ~~ 2799Secondarytuyeres pieces 12p02 supply Nm3/hr 5i3N2supply Nm3/hr 31 9LPGsupply Nm3lhr 12.8

Bottom tuyeres pieces 2p02 supply Nm3/hr i 97N2supply Nm3/hr 1ooLPGsuppiy Nmo/hr 4.6

Product'vit

Production rate tHM/daylm3 4602 supply Nm3lhr/m3 202

Table 6. Operational results of I ton furnace.

ItemsOff gas temp.Post combustion02 rate

Fuel rateProduction rate

Hot metal temp.

C in HM

unit

~~

alo

Nm/tHMk /tHMtHM/hr

ic

Wt-Qlo

Actua34015.8

1482471681454438

Calculated

40714.1

157246160i4644.23

liquid near the furnace wall is higher than that in theother zone because of the dripping through the highternperature zone in front of the primary tuyere. Thecal-

culated longitudinal distrlbutlon of temperature in thefurnace was good agreement with the observed one asshown in Fig. 12. Table 6 shows comparison of theoverall actual operational results with the calculated

ones. Thevalidity of the modelwasconfirmed from theseresults for the steady state simulation.

6.2. Results on the lOTonBOFType Test ConverterIn order to verify the validity of the model in respect

of scaling up the large scale furnace from the small scale

furnace, the simulation was performed on the actualoperational data of the 10 ton BOFtype test converter.The calculation wascarried out with a special emphasison the evaluation of dynamic response ability of themodel. Process parameters wasused the sameoneswhichwas used for the I ton furnace calculation. The opera-tional condition for the simulation Is shownin Table 7.

Theproductivity (t/d • m3) of the 10 ton furnace is about2.5 times as muchas that of I ton furnace as shownin

Table 7. The grids for the simulation are shown in

Fig. 13. The number and position of tuyeres weredeslgned based on the results of the I ton furnacesimulation. The numberof tuyeres were increased to

repress deflection of gas and solid flow in the circum-ferential and radial direction with the scale up. Fur-thermore, the primary and secondary tuyeres were in-

stalled at staggered position in the circumferential di-

rection as shown in Fig. 13 to avoid formation of re-

markable high temperature zone inside the furnace.

The computational region was treated a quarter of thefurnace considering its symmetry. The calculation pro-cedure was divided into two steps. At the first step,

upper(3rd.)

Secondarytu yeres

lower(2nd.)

Primarytuyere

Fig, 13. Grids of 10ton furnace.

the calculation was carried out to solve the initial

condition of the internal states in the furnace that

meansthe heating process of packedbedcoke which wascharged onto the bottom in the converter to form apacked bed of coke extending to just below the lowersecondary tuyeres during the coke packing operation. Inthe second step, on the basis of this initial conditionthe calculation for the scrap melting operation wasdynamically performed by the time difference At (= IOs)

according to the transition of the blowlng conditions asshownin Fig. 14. Thetransition of the operational results

calculated by the modelwaswell agreed with the actualoperational results as shown in the figure. From this

result the validity of the model was confirmed for theunsteady state simulation. Figure 15 shows the internal

behavior after 2hours from the start of the scrap meltingoperation calculated by the model. The internal statessuch as gas temperature and COprofile show threedimensional distribution and form non-uniform area.This behavior that flowery high temperature area wasformed in the circumferential direction wasalso observ-ed by a fiber scope equipped above the furnace. These

71 1 C 1999 ISIJ

ISIJ International, Vol. 39 (1999), No. 7

_ 800 1St

~ 600(~;~400~ 200

3rd I bIOW~)ff iBottom

2ndl I

IL l-\ !~l L

500~

2~ ;~ 300I

z- Ioo . :ist'

_ _ _ _ _ lBottom

I 3rd.L•L,,

1'_~f

50o c__ 1st

. I I l lI~ ~~ 30J 3rd. l i I~ Io ottom 2~nB. \

(QO ~ 600

o ~S~* 400

i st.

l I I- 200

~~ 20>.- ~ 15S:: J:: in1:, ~~ Iu~ 5O~

ool I ooCal. l l l o

OO

Oobs,o oo

~ 1700 CalE - 1600~~~)o, ~ 1500

i I lo

~: 1400 Obs. ooj l{o ci- 800c" E~) 600~G)-o

~ 400

l 1Cal Obs I l

,:: 50o l oo~jl lo o

8~~~ 30~!2 'E~o 10

o

Ca.

oI o

O(~Oobs o

o

1 2 3 4 5 6 7Time (h)

Fig. 14. Transition of the operational results calculated by the model for the actual operation of 10 ton furnace.

Post combustion

Vtuyere t IGasflow and Liquid flow and COgasSoiid flow andtemperature(~C) temperature(~C) concentration(vol,o/,)temperature(~C)

Fig. 15. Internal behavior of lOton furnace calculated by the present model.

indicate that there is a possibility for the numberandthe arrangement of tuyeres to have a great influence onthe operational results. However, as remarkable high

temperature zone does not exlst and scrap is melt stably

above the primary tuyeres as shownin this figure, it is

found that the present setting of tuyeres is applicable.

Thediscussion about three dimenslonal behavior is men-tioned above and further study is necessary whetherthe more suitable setting for the arrangement of the

secondary tuyeres is or not in order to achieve the morehigh heat efficiency with the best use of the sensitive heat

Table 8. Operational results of 10 ton furnace.

Items Unit Actual CalculatedOft gas temp. ~c 779 807Oft as volume Nm/hr 2744 2721Post combustion % 23 23Product'on ton 33 32Hot metal temp. ic 1440 1449C in HM olo 4.5 4.502 rate Nm/tHM 167 i 72Cokerate k /tHM 158 i51Ooal rate k /tHM 54 58LPGrate k /tHM 12 13Fuel rate k /tHM 224 222

C 1999 ISIJ 71 2

ISIJ International, Vol. 39 (1999), No. 7

Table 9. Calculation condition for 160 ton furnace.

Items 1Oton 80F 160 ton BOFFurnacedimensionDiameter m 18 5.6Hei ht m 3.6 8,0

nner volume m 7.4 197o erational conditionScra dameter mm 175Coke diameter mm 6002 su l Nm/hr/m 245Coal / I =t 02 () 0.82nd.02/1 st 02 () 0.8

Flametemp (~~) 2240Heat ioss coefticient (Mcal/m2/hr/~C) 8,0

Table lO. Effcct of the production scale on operationalresults

ItemsHeat loss Mcal/tHMFuei rate k /tHMPost combustion '/-

Off gas temp (~C)

Production rate tHM/da ImHot metal temp. (iC)

10 ton BOF13.7

18526

43746

i400

160 ton BOF3.4

17324

35447

1400

of the post combustion In case of using relatively lowheight reactor like the BOFtype furnace as the PSMprocess. Table 8 shows the comparison of overall op-erational results observed by the operation with thatcalculated ones during five hours from the start of themelting. Agood agreementwasalso obtained regardingthe overall heat and massba]ance as shownin this table.

Therefore, the scale up abillty of the mode]wasconfirmedfrom these computedresults.

7. Estimation of Operational Results on the CommercialSca]e PSMProcess

The calculation was performed to estimate the op-erational results of a commercial scale PSMprocessunder the following assumptions.l) The commercial scale plant was assumed 160 ton

heat size BOFtype converter.2) Typical operational data of the 10 ton furnace was

used as base data for the estimation.3) Steady state operation is performed.4) Scrap and coke are charged into the furnace with

separated charging.5) Burden distribution of the top of the furnace is uni-

form distribution.

6) Blowing conditions are not changed during the op-eration.

7) Heat loss coefficients (kcal/m2,h•'C) through the

furnace wall and the bottom is constant regardiess

with the furnace scale.

Table 9shows the dimension and ca]culation conditionfor the 160 ton BOF.The calculated results was shownin Table lO. The scrap me]ting ability of the commer-cial scale BOFwas still maintained stable under the

appropriate operational conditions as the sameway asthe I ton and 10 ton experimental furnaces even if thefurnace was scaled up to the 160ton BOF. It wasestimated that the fuel rate went downby the decreaseof heat loss rate that was reduced by the effect of the

lower ratio of surface area/internal volume than its ratio

of iOton furnace according to the scale up. From this

71 3

result the scrap meltlng tests using the lOton BOFtypefurnace wasconfirmed to be an useful methodto designthe PSMcommerclal plant.

8. Conclusions

The following results were obtained through the scrapmelting tests using the two furnaces (modified experi-menta] blast furnace of I ton pig per tap scale andmodified experimental BOFof 10 ton pig per tap scale)

by the PSMmethodand the numerical analysis by the

developed 3-dimensional mathematical simulation mod-el.

1) 1000/0 steel scrap can be melted using blast furnacecoke and pulverized coal.

2) The hot metal whose quality is equal to that ofblast furnace can be produced with high carburiza-tion (~4 masso/o C) and high desulfurization (~~O.03masso/oS).

3) High heat efficiency(~800/0) and high productivity(50 t/d • m3) were achievable.

4) Hlgh iron yleld (~99 masso/oFe) and low refractoryerosion (~0.5mm/charge) can be achieved by the

packed scrap and coke.5) The analysis of three dimensional and dynamic

behavior of the process becamepossible and theexperimental results using the two furnaces were welljustified by the model.

6) A basic design of commercial scaie plant was also

performed by the use of the model. As a result, it

wasconfirmed that the PSMmethodusing commer-cial scale BOFwasuseful as a scrap melting process.

Nomenclature

X: coordinatet: time

Dj: effective diffusion coefficient of gaskj: thermal conductlvity ofphase.j

Cpj : meanheat capacity of phase.j

uj,,, : heat transfer coefficient between phase ,j andphase m

Rate,, : rate of reaction nAH,, : heat of reaction n

Mi : moiecular weight of species iaji,, : stoichlometric constant of species i of phase .f

at reactlon naj,,, : contact area between phasej and phase mdp : Particle diameterTj: temperature of phasejP: gas pressureU: velocity vectorB: klnematic model's parameterej : volume fraction of phasejpj : density of phase.j

~ji : massfraction of species i in phase.j

,1j,, : distribution ratio to phasej of the heat of re-action n

p: viscosity of gasip : shape factor

Subscripts

l : Iiquid, s: solid, g: gas and i. ./, k: coordinates

vector

O1999 ISIJ

l)

2)

3)

4)

5)

6)

7)

8)

9)

lO)

ll)

12)

l3)

l4)

l5)

ISIJ International, Vol. 39

REFERENCEST. Yamamoto:Symp. Recent Trends of Newiron ProductionProcess. ISIJ. Tokyo. (1996). I l.

K. Musaka: I14, 115th Nishlyama Memorial Seminar ISIJ,

Tokyo. (1986), 173.

M. Dietlin, J. J. Klein, M. Helleisen and F. Leclercq: Rev. Metal!.

Cah. I,nf: Tech.. 81 (1984), 734.

S. Okamura, H. Nakajima, K. Marukawa, M. Ane7.aki, Y.Tozaki, A. Mori, K. Katou, K. Ichihara: Tc'tsu-to-Haga'li, 71

(1985), 1787.

T. Yamamoto.Y. Ujisawa, H. Ishida, T. Mori, H. YamaokaandK. Marukawa:Metallurgical Process for Early TwentyCentury,2. TMS,(1994). 843.

H. Yamaokaand Y. Kamei: Tc'Isti-to-Hagcinb, 79 (1993), 139.

S. Ogura: 114, I15th Nishiyama Memorial Seminar ISIJ. Tokyo,(1986), 8.

H Ishida, K. Marukawa.S. Anezaki. T. Hirata and S. Shiode:

Process Technology Conf. Proc. 7, ISS-AIME, (1988), 233.

J-P Birat. R6myNicdle. M Dietlin. M. Helleisen. J. J. Klein

and M Schneider: CAMP-ISIJ. 6(1993), 1024.

Y. Ujisawa. H. Ishida, K. Takatani. T. Yamamotoand H.Yamaoka:CAMP-!SIJ, 7(1994), 30.

Y. Ujisawa. H. Ishida, K. Takatani, T. Yamamotoand H.Yamaoka:CAMP-ISIJ,7(1994), I140.

J. Yagi and I.Muchi: T,'a,Is. I,'on S!ee! Insl. Jpn., 10(1970). 39'-.

R. Wartman:Stah! Eise'l, 95 (1975). 442.

MHatanoand K. Kurita: Trclns. I,'on Stee/ Ins!. J/;n., 22 (1982).

448.

M. Hatano, K. Kurita, H. Yamaokaand T. Yokoi: Trans. Iron

16)

l7)

l8)

19)

20)

21)

22)

23),~4)

25)

26)

27)

28)

29)

30)

31)32')

33)

(1999), No. 7

Stee/ his/. Jpn., 25 (1985), 933.

T. Sugiyamaand M. Sugata: Sei!e!su-Ke,7/~'yu. 325 (1987), 34.

S. Ergun: Chem.Eng. Prog.. 48 (1952). 89.

RM.NeddermanandU. Tuzun: Po]t'c!e,' Tecllno!., 22 (1979), 243.

KagakuKohgakuBinran Kaitei 5han, Maruzen, Tokyo. (1998),

988.

K. Takatani. Y. Ujisawa andH. Yamaoka:CAMP-ISIJ,2(1989),

l030.

H. Yam'aokaand Y. Kamei: Teisu-to-Hagand, 74 (1988), 2254.I. Ohnakaand T. Fukusako: Trans lr0'7 Stee! h!st. Jpn., 17 (1977),

410.

Y. Iwanagaand K. Takatani: J. Fue! Soc. Jpn., 68 (1989), No. 4.

K. Takatani and Y. Iwanaga: Telsti-lo-Hagane, 75 (1989), 594.

W. E. Ranz: Cllenl. Eng. P,'og.. 48 (1952), 247.

W. H. Gledt: Principles of Engineering Heat Transfer, transl, byS. Yokoborl and O. Kuga, Maruzen, Tokyo, (1957),146 and 281.

KagakuKohgakuBinran Kaitei 5han, Maruzen, Tokyo, (1998),

355.

KagakuKohgakuBinran Kaitei 4han. Maruzen. Tokyo, (1978),

58.

K. Onda, H. Takeuchi and K. Koyama:Kagaku-Kougaku,31(1967). 127.

J. Umada,H. Shinohara and H. Tsubakihara: Kagaku-Kougakti,27 (1963), 978.

K. Takatani: Te/su-to-Hagclni, 74 (1988), 1546.

C. W. Hirt. B. D. Nichls 'and N. C. Romero:SOLA-ANumericalSolution Algorithm for Transient Fluid Flows. Los AlamosSci.

Lab, Los Alamos, (1975). LA-5852.

P J Roachc: Computational Fluid Dynamics, HermoraPub-lishers, (1976)

o 1999 ISIJ 71 4