and slag in the submerged-arc furnace - saimm

The mode of current transfer between electrode 

. 

and slag in the submerged-arc furnace 

SYNOPSIS 

by W. P. CHANNON*, B.Sc.(Eng.)(Student Member), 

R. C. URQUHARTt, M.Sc.(Eng.), ;'h.D.(Visitor), 

and D. D. HOWAT*, B.Sc., Ph.D.(Fellow) 

An account is given of a laboratory experiment in which the resistivity of a primary slag like those produced in the 

manufacture of high-carbon ferrochromium was measured at current densities typical of operation in a submergedarc 

furnace. It was shown that the current flow from graphite electrode to molten slag obeys Ohm's Law when 

electrode current densities are less than 12 A/cm". At higher current densities, arcing occurs from the electrode 

tip to the slag, causing an imbalance in the rate of heating in the furnace burden. 

SAMEVATTING 

In the last twenty years several 

dramatic changes have occurred in 

the field of high-temperature smelting. 

One of these changes was the 

introduction of electric power for 

the generation of high temperatures, 

in place of fossil fuels. The 

ever-increasing shortage of coking 

coal has resulted in the submergedarc 

furnace becoming a valuable 

metallurgical tool for the production 

of ferro-alloys. The energy requirements 

for the production of ferroalloys 

vary from 2 to 10 MWh per 

tonne of alloy produced, and the 

cost of electric power is sufficiently 

high for economy of operation to 

be essential. 

It has been maintained that the 

production from electric reduction 

furnaces is a direct function of the 

resistance of the burden in the 

furnace1. The heating ina ferroalloy 

furnace is not uniform because 

the constituents of the burden 

undergo physical and chemical 

alteration as they progress down 

through the furnace. This has motivated 

studies on the variation, with 

temperature, of the electrical resistivity 

of the individual burden components, 

of the mixed burden components, 

and of the molten slag1-7. 

In addition, it has been shown that 

the major proportion of the power 

in a submerged-arc furnace is dissi- 

*University of the Witwatersrand. 

tNIM Pyrometallurgical Research 

Group, University of the Witwatersrand. 

4 AUGUST 1974 

Daar word verslag gedoen ocr 'n laboratoriumeksperiment waarin die spesifieke weerstand van 'n primere slak 

secs dill wat in die vervaardiging van koolryke ferrochroom verkry word, gemeet is by stroomdigthede wat tipies is 

van werking in 'n dampelboogoond. Daar is getoon dat die stroomvloei van die grafietelektrode na die gesmelte 

slak die Wet van Ohm gehoorsaam wanneer die elektrode se stroomdigtheid minder as 12 A/cm" is. By hoer stroomdigthede 

vind daar oorslag van die elektrodepunt na die slak plaas wat 'n wanbalans in die verhittingstempo van die 

oondlading veroorsaak. 

pated in the region of the tips of the 

electrodes8. However, a certain 

amount of confusion seems to exist 

about the manner in which this 

power is dissipated. Some authors9,1o 

claim that the heating is mainly 

resistive, whereas others favour the 

theory that the main heat is generated 

in the arc zone at the tips of 

the electrodesll. 

EXPERIMENTAL 

INVESTIGATION 

In order to investigate the mechanism 

of current transfer from 

electrode to molten slag, a simple 

experiment was designed to measure 

the resistivity of the slag at current 

densities typical of operation in a 

submerged-arc furnace. In a singlephase 

circuit, current at 50 Hz 

frequency was made to flow via a 

graphite electrode through a primary 

slag typical of those produced in the 

manufacture of high-carbon ferrochromium 

(22 per cent MgO, 28 per 

cent A12O3' 50 per cent SiO2, saturated 

in carbon). The experimental 

apparatus is shown in Fig. 1. The 

current density on the graphite 

electrode could be varied between 

0 and 15 Afcm2, the graphite crucible 

acting as the second electrode. 

Experiments were performed in a 

molybdenum-wound resistance furnace 

at slag temperatures in the 

region of 1500 DC. The voltage across 

and the current flowing through the 

molten slag were observed on an 

oscilloscope screen. The graphite 

electrode and crucible were pro- 

tected against oxidation by purging 

of the system with argon gas. At 

electrode current densities of less 

than 12 Afcm2, it was found that the 

oscilloscope trace of the current 

flow pattern was sinusoidal and 

Ohm's Law obtained. Figs. 2 (a) and 

(b) show the form of the oscilloscope 

traces for the potential across the 

slag and the current flowing through 

it. However, it should be remembered 

that the magnitude of the slag 

resistivity under these 'submergedarc 

furnace-like' conditions of high 

current density and low frequency is 

higher than the absolute resistivity 

as determined in fundamental studies 

in the laboratory 6. 

At current densities in excess of 

12 Afcm2, the pattern of current 

flow as shown on the oscilloscope 

was characterized by the occurrence 

of unstable arcing. The voltage and 

current waveforms typical of those 

obtained at current densities greater 

than 12 Afcm2 are illustrated in 

Figs. 2 (c) and (d). As all the experiments 

were done with the 

graphite electrode immersed in the 

molten slag, the occurrence of arcing 

was at first puzzling. However, with 

the high current densities at the tip 

of the electrode, extreme localized 

heating occurs that leads to an 

increase in the rate of reaction 

between the carbon electrode and 

the slag, resulting in the formation 

of gaseous products, CO, and SiO. 

The rate of reduction of SiO2 to SiO 

by carbon is rapid at temperatures 

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY

Argon 

2 

3 

4 

5 

Cl 

7 

JL 

Fig. I-The apparatus used in the ex. 

perimental determinations 

I Graphite rod 

2 Sindanyo collar 

J Graphite spacer ring 

4 Electrode 

5 Inner alumina tube 

6 Furnace heating coils 

7 Outer alumina tube 

8 Graphite crucible 

9 Graphite support rod 

10 Alumina thermocouple sheath 

II Sindanyo disc 

12 Support collar 

greater than 1500oC12, and confirmation 

of this mechanism was 

obtained by the presence of a precipitated 

film of silica in the upper 

regions of the experimental furnace. 

The electric circuit for the system 

under arcing conditions is shown in 

Fig. 3. As shown in Fig. 2 (c), the 

voltage waveform represents the 

total voltage drop across the series 

combination of the arc and slag 

resistance, that is, the sum of the 

square wave voltage drop across the 

arc and the sine wave voltage drop 

across the slag resistance. The 

9 

(J) 

..., 

.-j 

D 

:::0 

50 

40 

30 

20 

10 

-10 

-20 

-30 

-40 

8 

6 

4 

2 

(J) 

CD 

1-1 

CD 

0- 

E 

et: 

- 2 

- 4 

- 6 

(J) 

..., 

.-j 

D 

:::0 

(J) 

CD 

1-1 

CD 

0- 

E 

et: 

30 

20 

10 

-10 

-20 

6 

- 6 

(a) 

(c) 

(d) 

seconds 

Fig. 2-Waveforms obtained for conduction experiments in the slag medium 

(a) Sinusoidal voltage drop across slag resistance. 

(b) Sinusoidal current nowing through the slag. The current lags the voltage by 

18°. 

(c) Voltage drop across the arc and the slag resistance. 

(d) Arcing current and theoretically predicted arcing current. 

voltage trace is characterized by a 

sharp rise to the peak voltage required 

to re-ignite the arc (et). As 

soon as the arc strikes, current flow 

is re. established and the voltage 

drops to the arc voltage (ea), which 

remains constant for the remainder 

of the half cycle. The sinusoidal cap 

on the voltage trace represents the 

voltage drop across the slag resistance. 

It is of interest at this point to 

call attention to the fact that, in 

Figs. 2 (c) and (d), the ignition and 

arc voltage, as well as the current, 

differ in each half cycle. This effect 

is primarily due to the differences in 

cathode behaviour of the graphite 

electrode and the slag surface as the 

polarity reverses with the alternat. 

ing current. During the second halfcycle, 

when the graphite electrode 

was the cathode, the arc conductivity 

was superior as shown by the lower 

arc voltage and higher current. This 

is in agreement with a similar finding 

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY AUGUST 1974 5 

2

_I 

Graphite 

Electrode 

- 

.~. ~~ 

(a) 

-::-:::-- ~~-~= 

?- --- :. --- .- -r~olten -- 

51ag 

5io co e a 

~ 

- .~ It~~ 

- - ...----- .--- 

--- - -- - - 

C + 5i02 - CO+ 5io 

(b) 

Il.. 

es 

Av 

r s 

f:aood~ 

x r 

am "O 

x r 

~ ~: 

Fig. 3-A schematic diagram illustrating the two possible means of conduction 

(a) Ohmic conduction as shown by the equivalent circuit and the sinusoidal voltage 

drop across the slag resistance. 

(b) The equivalent circuit under conditions of arcing and the corresponding 

voltage drop across the arc and slag resistance. 

by Schwabe13 and Bowman et al.14 

concerning open-arc steel-melting 

furnaces. 

The equation describing the arcing 

circuit of Fig. 3 is: 

and 

4e+i R+ea=Em sin (wt-cp), 

where wt ~ cp 

cp=arc ignition angle with 

respect to the supplyvoltage 

zero crossing, 

. 

-l. 

ei 

SIn .,.,= 

Em ' 

L=the inductance of the 

circuit, and 

R=the resistance of the slag 

and circuit. 

In the case of open-arc steelmaking 

furnaces, the voltage drop 

across the bus bar resistance (ir) is 

small compared with ea, and so this 

term is usually neglected in the 

above equation. However, in the 

submerged-arc equivalent, the voltage 

drop across the slag resistance 

is not insignificant. A solution of the 

above differential equation has been 

given as :15 

i= ~m {sin (wt-cp)-sin (cp-a) 

[ - R

12 A/cm2 determined as the limit for 

ohmic conduction in this study. It is 

expected that this value would be 

larger than that for a production 

unit owing to higher heat losses from 

the laboratory system. However, 

areas of high current density may 

occur at the tip of the electrode, and 

consequently an arc might be struck 

from the electrode to the slag. If 

an arc strikes from the tip of one 

electrode in a three-electrode submerged-arc 

furnace, then the resistance 

of this phase will be higher 

than that of the remaining two 

phases, resulting in an unbalanced 

three-phase load. This imbalance in 

the three-phase load will be further 

aggravated by the difference in arc 

resistance between the positive and 

negative cycle. The power in the 

arcing phase will be lower than in 

the other two phases, and, as a 

consequence, the total power dissipated 

in the furnace will decrease. 

Although an imbalance in the threephase 

load is not detrimental to the 

workings of the furnace transformer, 

it will result in uneven heating in 

the furnace burden and this may 

cause difficulties in operation. On 

furnaces producing high-carbon 

ferrochromium, it has been observed 

that fluctuations of the voltage and 

current meters may occur for short 

periods following a furnace tap. 

This is probably caused by burden 

falling away from the tip of the 

electrode and an arc being struck as a 

Iesult. Reactance meters, or similar 

devices that can detect changes in 

phase angle, can be used to warn of 

the occurrence of arcing. The operator's 

reaction is to decrease the electrode-to-electrode 

voltage and/or to 

raise the electrode slightly. In short, a 

decrease in the voltage gradient will 

serve to extinguish the arc. 

It seems that, in the design of 

submerged-arc furnaces, the current 

density in the electrode is chosen so 

that arcing in the furnace will be 

avoided. In Table I it appears that 

the permissible current density increases 

as the endothermic nature of 

the reduction reactions increase. 

The lower limit appears to occur for 

the smelting of copper-nickel concentrates 

in which there is virtually 

no endothermic heat of reaction. 

Product 

TABLE I 

FeCrSi (Single step) 

FeCr (High carbon) 

FeMn (High carbon) 

CuNi matte. . . . . 

CONCLUSIONS 

Current density 

A/cm2 

6,5 

5,0 

4,0 

1,5 

The laboratory experiment has 

shown that current flow from a 

graphite electrode to a molten slag 

obeys Ohm's Law at electrode 

current densities of less than 12 

A/cm2. With current densities in 

excess of 12 A/cm2, arcing occurs 

from the electrode tip to the slag. 

Arcing is not desirable in furnaces 

designed for resistance heating because 

it results in a decrease in 

power delivered to the furnace at a 

given voltage. The striking of an 

arc from one electrode effectively increases 

the resistance of that phase 

with respect to the others, causing 

an imbalance in the rate of heating 

in the furnace burden. Since the 

presence of an inductive load in the 

circuit, with its effect of decreasing 

the power factor, tends to create 

conditions for stable arcing, large 

smelting units with their inherently 

low power factors may be subject to 

some arcing problems. 

ACKNOWLEDGEMENT 

The authors acknowledge the assistance 

of Mr R. de Kock, of the 

Instruments Division ofthe National 

Institute for Metallurgy, who helped 

with the recording of the oscilloscope 

waveforms. 

REFERENCES 

1. COR:aeA DA SILVEIRA, R. Electrical 

and metallurgical reasons for some 

charges employed in electrical reduction 

furnaces producing ferro-alloys. 

JOURNAL OF THE SOUTH AFRICAN INSTITUTE OF MINING AND METALLURGY 

Metalllurgia A.B.M., vo!. 25, no. 136. 

1969. pp. 167.178. 

2. VOLKERT, G. Influence of slag composition 

on the operational characteristics 

of an electric reduction furnace. 

Chemie Ingenieur Technik, vo!. 42, 

no. 4. 1970. pp. 218-221. 

3. LIUTIKOV, L. A., and TsYLEv, L. M. 

Viscosity and electrical conductivity 

of MgO-Al.O3-SiO. melts. Izv. Akad. 

Nauk. S.S.S.R. Old. Tekn. Nauk. 

MetaU.i.Gorn. Delo, no. 1. 1963. 

pp. 41-52. 

4. LIUTIKOV, L. A., and TsYLEv, L. M. 

Effect of chromium oxide on the 

viscosity and electrical conductivity 

of melts of the MgO-Al.O3-SiO. 

system. Ibid., no. 2. 1963. pp. 59-66. 

5. OSSIN, D. 1., et al. Liquidus temperatures, 

viscosities and electrical conductivities 

of lime-containing slags 

produced during the smelting of 

high-carbon ferrochromium and ferrochromium-silicide 

alloys. 29th Electrical 

Furnace Conference, 1971. pp. 

94-101. 

6. RENNIE, M. S., et al. The effects of 

chromium oxide, iron oxide, and 

calcium oxide on the liquidus temperatures, 

viscosities and electrical 

conductivities of slags in the system 

MgO-Al.O3-SiO.. J. S.Afr. Inst. Min. 

MetaU., vo!. 73, no. 1. Aug. 1972. 

pp. 1-9. 

7. DOWNING, J. H., and URBAN, L. 

Electrical conduction in submergedarc 

furnaces. J. Metals, Mar. 1966. 

pp. 337-344. 

8. URQUHART, R. C., et al. The dissipation 

of e]ectrical power in the burden 

of a submerged-arc furnace. 31st Electric 

Furnace Conference, 1973. 

9. ELYUTIN, V. P., et al. Production of 

ferro-alloys. ElectrometaUurgy, 2nd 

edition, Israel Program for Scientific 

Translations. Jerusalem, 1957. 

10. PASCHKIS, V., and PERSSON, J. 

Industrial electric furnaces and appliances. 

New York, Interscience Publishers, 

1960. 

11. VOLKERT, G., and FRANK, K.-D. 

MetaUurgie der ferrolegierungen. 2. 

Neubearbeitete una erueiterte auflage. 

Berlin, Springer-Verlag, 1972. 

12. KLINGER, N., et al. Reactions between 

silica and graphite. J. Amer. Cer. 

Soc., vo!. 49. 1966. pp. 369-375. 

13. SCHWABE,W. E. Measuring problems 

and techniques at A-C furnace arcs. 

J. Electchem. Soc., vo!. 101. 1954. 

pp. 554-559. 

14. BOWMAN, B., et al. The physics of 

high current arcs. J. Iron Steel Inst., 

Jun. 1969. pp. 798-805. 

15. LEES, L. H., and LUXAT, J. C. 

Requirements for stable A.C. electricarc 

discharges. Electronics Letters, 

vo!. 6, no. 9. Apr. 30, 1970. pp. 

285-286. 

16. SELMER-OLSEN, S. Trends in ferroalloy 

production. J. S.Afr. Inst. Min. 

MetaU., May 1971. pp. 210-214. 

AUGUST 1974 7

and slag in the submerged-arc furnace - saimm

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?