60th Electric Furnace Conference Proceedings, Vol. 60, Warrendale, PA: Iron & Steel Society, 2002. pp.

425-437

THE APPLICATION AND SCALE.UP OF A.C. AND D.C.

SMELTING FTIRNACES FOR FERRO-ALLOYS

Nicholas A Barcza, Ian J Barker, Myles S Renaie, & Paul J Brereton-Stiles

Mintek
Private Bag X 3015
Randburg 2125
' South Africa.
Tel.: +27-11-7094111
E-mai1s: [email protected]; [email protected], mvlesr@mintek co.za, oaulbs@mintek co za;

Key Words: Fenoalloys, Submerged-Arc Fumace, PIasma lumace,

Burden Conductivity, Limits of Operation, Scale-up

INTRODUCTION

The prices of feno-alloys, as with most commodities, have fallen significantly over the past tlventy years.
Thisias forced many plants to close or cut costs drarnatically to survive. Technology has played a cntical
role in this cost reduction. lmproved feed material preparation, larger and more efficient submerged arc
furnaces, changes in post-taphoG practices, and advanced process control have all had majol impacts on the
industry. Th"ie would seem to be advantages of scale by going larger still, but to achieve this, a different
upprou"h *.ill be needed, as any further advances with conventional technology would seem to be rather
limited.

Submerged-arc firmaces continue to be used for feno-alloy production These fumaces typically use three-
phase Jtemating-current power, fed in through three electrodes arranged in a triangular configuration.
There wolld seem to be advantages ofscale with these fumac€ operations, but various limits of capacity are
now being encountered with thi conventional technology, and these are becoming a hindrance on the
constructiin of larger units. In the case of high-carbon ferromanganese, for example, it would seem to be
very difficglt, if nit almost impossible, to achieve operation at frwer ievels above aboul 40. to 50 MWt.
Wiih silicon, which also has a rilatively lorv electrical resistance, the limit would seem to be about 25 to 30
MW, while with high-carbon ferrochromium, llhich has a higher resistance, the present limit rvould seem to
be up to about ZO tvfw. By comparison, the largest steel-making arc fumaces are curently of the order of
80 to 100 MW, and this figure would seem to be still rising.

ln the case of high-carbon ferromanganese, this constlaint on firmace size means that it is difficult to get a
yearly productiol per fumace of more than about 140 to 160 thousand metric tonnes pel amum When this
is comiared to the rate of production of pig iron from a modem blast fumace, which is tlpically
of the order
of 2 to 3 million metric tonnes per arLnum, but in some cases is up to 5 million tomes per annum, it is
evident that the capacities of even the largest submerged-arc fumaces are small in comparison. Questions
arise as to rvhv this is the case, and whether ferro-alloy fumaces could or should be made any larger.

A1<
2002 Electric Fumace Conference Proceedings
In this paper we illustrate how some of the constraints on conventional submerged-arc fumaces can rn
principll be rela,red through the use of d.c. power. In addition we indicate the potential benefits of d.c.
fumace technology to the ferroalloys industry.

power
It is convenient conceptually to split the characteristic curves of a ferro-alloy process into those ofthe
supply and those of tire furnace. From the practical perspective of the power suPPlY, it is possible to get
curve
an a.c. or a d.c. power supply that can deliver power in almost any region of this characteristic
"itt
gaph "i where the fumaie may require it However, it has been. fourd that as the process is scaled up, so the
operatlng polnt seems to move lnto areas of these characteristic cuwes that inherently
incur operational
pioblemi from tne fumace side. Two particular problems that have been identified are the current-carrying
ofthe electrodes. In theory, d c power
iapacity of the electrodes and the inteiaction effect in the control
snoUa-netp to alleviate both of these problems, as this paper will attempt to explain in more detail'
However, d.c. has some problems of its orn, and there may be still more hidden
problems that will become
paper will discuss some ofthese
evident only once these more obvious gonstraints have been removed. This
possibilities.

TEE FURNACE CIRCUIT

Consider a section through a fumace around one electrode, as shown schematically

in Fig. 1, with an
ignore the issue of how this supply is brougltt in,
electrical supply to that one electrode. At this stage, let us
floor of the heartlu or via a retum path
except to ,uy ttrut it can be done either through a connection to the
thoush other electrodes in the same fumace.

Electrode Magnelc
inductance
Magnetic field
around
etecIIooe

Burden Source of
electrical
power
Hot
zone

Current
paths

(a) Schematic ofthe physical arrangement (b) Equivalent circuit

Fig. 1: Conduction tlrough one electrode ofthe fumace, and its equivalent circuit.
the hearth
In this fimace circuit, the power supply creates a voltage between the top of the electrode and
to the bearth The
This voltage drives the current to flow through the electrode and then through the bulden
ofthe burden beneath the electrode, and there ts
majority oI the electrical power is dissipated tn the hot zone
the hearth. The cunent flowing
tt eiefoie a significant voltage drop betueen the tip of the electrode and

2002 Electric Furnace Conference Proceedings

 reveals itself as an
around the circuit creates a sigrificant magnetic field around the conductors, and this
electro-magnetic inductance in each phase. Any a.c. current flow through this inductance cleates an
is of little importance as it
additional i.c.-voltage drop. However, in the case ofd.c. power, the inductance
does not produce a steady-state voltage drop

Characteristic curves
through the
A graph can be drauryr ofthe power dissipated beneath this electrode against the current.flowing
power supply
oa", as shown in Fig. 2. The process itself will have one characteristic cuwe while the
intelsect
wrll havi another as indrcated. Tie operating point will be
"tel"t the point at which these cuwes
(provided that they do, and that the point ofintersection is a stable operating point)

characteristic culve, as
The gently upwards-curving line in Fig. 2 is fairly typical of the shape of a furnace
will 6e discuised in more ditail belowl The inverted-U-shaped curves in Fig. 2 are tlpical ofan
a c. supply

witt a -agnetic inductance in series With the fumace load, and fed from a fixed voltage source lf the
voltage oithe supply is changed while the remainder of the circuit is kept fixed, a different curve ts

obtained. The other two rnverted-U shaped lines in Fig. 2 are two examples of this Hence, by tapping the
the furnace cuwe Thus, in
transformer voltage up or down, it is poisible to move ihe operating point along
graph where it may be
theory, it is poss-ible'to provide power in almost any re_glol of the characteristic
.qrli"a nurthermore, by tapping the supply voltage up, both the power and the current increase
Characteristic curve
,z--'-- of fumace

ul

Characteristic
cuwe of power
supply

Electtode curteld

Fig. 2: An example of the characteristic curves for the fumace and the power supply

Trends in furnace sizes

Thefeno.alloysindustryemploysagreatnumberoffumaces'operatingonmanydifferentproducts'The
a characteristic power-versus-current
operating points of these fumacis shJw that each distinct process has
5%o fenosilicon, silicorL and two
il;. ;G. 3 shows some obsewed trends for ferrosilicomangan ese, 7 prelTously- -'
types offenochromium processes Related trends have been documented

the region where any

An extrapolation of the trend lines beyond the ranges shorvn in Figure 3 indicates
future expansron in the size of fumaces will take place'

2002 Electric Furnace Conference Proceedings ^11

 75 % FeSi
=
=
3ro
E
F

120 160

Electrode current, kA
Fig. 3: Trends for five different ferro-alloy processes, on a graph oftotal power vs. electrode current.

Constraints on the Operation

The concept of a multiplicity of constraints has been described by De Waal et a[.6 lt atso applies to the issue
of scaling up of a ferro-alloy furnace. ln particular, when some factor that is limiting the process is
extended, thi operation is often unable to make full.use ofthis extension, because the operating point then
encounters another different limitation.

The followrng five factors have been suggested as being the major constraints on ttre maximum size of
present submerged-arc fumaces :
1. The current-carrying capacity of Soderberg electrodes,
2. The resistance ofthe burden,
3. The reactance ofthe porver supply circuit,
4. The gas permeability of the burden, and
5- The energy generated by the combustion ofthe fumace off-gases.
Of these, the reactance of the power supply is probably the main constraint. The cunent-carrying capacity
of the electrodes also appears to be a major limitation, but in reality it may be the reactance problem in
another form that appears as a limit on the electrode size. The resistance of the bwden is in most cases
essentially fixed, but there are some variations that can be considered. The gas permeability and the
combustiln ofthe off-gases may be further complications, but these can possibly be accommodated.

CHAR{CTERISTIC CUR1IES OF TITE FURNACE PROCESS

Denote the current through the electrode by i, and the power dissipated in the furnace by this current as P6
(ITB:this may not be the iame as the total powerPr in the furnace as a whole). Ifthe conduction path ofthe
current through the fumace behaves as a resistor ofresistance Rs, then the following equation would relate
the power to the curent:

Pa : .Rr
i2 (1)

On the other hand, the power-dissipating element may act more like an arc. Arcs are complex phenomena
electrically, but from a iimplistic vieupoint, the arc tends to behave somelvhat less like a resistor and more

428 2002 Electric Furnace Conference Proceedings

like a constant-voltage device. In this case, if this voltage is %, the followrng equation would relate the
power to the curent:

Pr = i.v, (2)

Based on actual data collected from fumaces, Westly proposed the C3 factor as another possible formula to
describe the characteristic curves of the fumaceo. This relationship can be re-written for a single elechode
as:

Ps: its l13 " crls1 (3)

(It is evident that the functional form of equation (3) is a compromise between equations (1) and (2)).

The above three relationships all indicate that:

.the greater the current, the greater the power, and
o at a given current, the larger ihe magnitude of the power-dissipating el€ment (i.e. the higher the
resistance, or arc voltage or reciprocal of Cj), the more power.

Accordingly, to raise the power dissipation in a fumace, one should look at ways to both increase the
current in the electrode, and ilcrease the magnitude of the power-drssipating element (e.g. the resistance)
between the electrode and the hearth.

Limits on Current per Electrode

The dominant factor that determines the maximum current is the diameter of the electrode. Other factors,
such as the quality of the paste, the design of the casings, the rate of slipping, and the operation of the
fumace, also have an influence, but are minor by comparison. Hence to increase the cimenl-carrying
capacity, it is basically necessary to increase the electrode diameter.

Despite all attempts to improve the electode technology over the years, Andreae's original formula for the
limrting cunent in a Soderberg electrode is still usedZ. This can be converted to metric units and re-
anangJd, to relate the limiting current 1r in kA to the diameter ofthe electrode D6 in metres, as follows:

Is = JJ x Puts (4)

At present, the Iargest sized Soderberg electrodes used in submerged-arc fumaces are tlpically around 1.7 to
1.9 metres in diameter. From equation (4), it can be sho*.n that these conespond to current capacities of
aboti 122 to 144 kA respectively. However, it is questionable whether this formula can still be used to
predict the capacities ofelectrodes that may be larger than about 2.0 metres in diameter, as it has never been
tested or verified in this region.

The main consequence ofexceeding the current-carrylng capacrty of a given size of electrode is that the risk
of an electrode break increases sigrrificantly, although excessive current is not the only factor that can cause
an electrode break. If an electrode breaks, not only does it affect the production from the fumace, but also it
may upset the metallurgy and so produce off-specificahon material. It would appear that larger electrodes
u.r rno." susceptible to breaking. Furthermore, the upset to the operation may also kigger another break,
i
and so lead to vicious cycle of electrode breaks that is difficult to recover from, but here the interaction
effect (as is described in more detail below) probably also plays a role.

We suggest that there are fwo direct reasons why a larger Soderberg electrode tends to give more trouble
than a smaller one, viz.:
1. The strength-to-weight ratio deteriorates as the electrode is scaled up.
2. The skin effect distorts the distribution of the current flowing in the electrode'

2002 Electric Fumace Conference Proceedings 429

Strength-to-weight ratio - The weight of the electrode rises by the cube of a Lineal dimension (e.g. the
ratio falls
alameter), while the strength rises byihe square of this dimension That is, the strength-to-weight
in propoiion to the recifrocal of the dimension as the electrode gets larger. . Hence. electrodes become
problem'
relatively weaker as they get bigger. It would appear that there is no simple solution to this

Skin effect - The skin effect refers to the tendency of a.c. current to keep to the outside ofa conductor'
It
does not occur with d.c. current. It is caused by voltages induced within the conductor by the fluctuating
magnetic field associated with the current flowing in the same conductor.
(This effect should not be
so-called 'motor
conlused with the forces on a current-carrying conductor in a magnetic field, as in the
cunent flow penetrates below the surface of
rule'). The skin depth is a measure of tne depih to which this
the conductor. This skin depth d is given by:

(5)

where p is the specific resistivity of the conductor, /is the frequency of the a.c. current in Hertz, /'{, is the
of the conductor' By. substituting
-"grr"ii" p"t-.ubitity of fiee siace, and p. is the relative permeability the
typlcal vaiues for these parameiers, it can be sho$n that for Ssderberg carbon and pre-baked carbon,
i<in thickness is about 0.3 to 0.4 metres, while for graphite it is about 0 15 metres

For any electrode where the radius is larger than the skin thickness, the skin effect will be significant.
Fig. 4
figure
shows some calculated current densities=across Soderberg electrodes for a range of diameters. This
assumes that all the relevant physical properties are uniform across the electrode, and
is based on the tpical
properties in the carbon stump ofthe eteitrode. Although the cuwes obviously do not strictly apply to the
the
tuking ,on" of the electrodg they do indicate that there may be problems rvith the baking towards
centre"s of the larger electrodes because of the deficit of current in these centre regions. It is evident from
Fig. 4 how the sk]n effect plays an iocreasing role as the diameters q.t lgtg"l: Also. because
the current is
thJn concentrated in the skin layer at the surface, the heating from the l' of
R dissipation power ls greater'
and so the elechode gets slightly hotter relative to the corresponding d'c case

R.dial posltion, melre!

Hz.
Fig. 4. current densities across a Soderberg electrode for various electrode sizes, at 50
(The total electrode current for each diameter is specified from equation (4) )

430 2002 Electric Fumace Conference Proceedings

With d.c. power the skin effect do€s not occur (as can be shown by substituting/= 0 in equation (5)), which
indicates tbat large electrodes may tend to behave better in a d.c. fumace. Conversely, smaller diameter
electrodes would behave similarly using a.c. and d.c. as their current densities would be uniform in both
cases.

Graphite electrodes can carry aLmost double the current of Soderberg electrodes of the same diameter, and
henie are a definite altemative. A comparison of power input using Soderberg and gaphite electrodes for
ferrochrome and ferromanganese production in a submerged-arc furnaces shows that with graphite
electrodes, roughly double the power can be achieved at the same current. At present, the largest graphite
electrodes are 0.8 metres in diameter (capable of carrying 140kA), but if the demand was there, larger
diameters could perhaps be manufactured- It may be possible to set oIf the high cost of graphite electrodes
against the productivity increases that would ensue. Probably the greatest challenge for the
"on.id"rubl"
olerator of a feno-alloy tumace with Soderberg electrodes is to balance the baking-slipping rates of the
eiectrodes, as required 10 maintain electrode lengths. Pre-baked and graphite electrodes do not have this
limitation, because they can be slipped as much or as Little as required.

In conventional a.c. submerged-arc fumaces at present, the electrode diameters are only one of many
constraints on fumace size. Other constraining factors, in particular the interaction effect (as explained
below), are probably more critical. However, if d.c. porver were used, then it is very possible tlat the
electrodes would be cdlled upon to carry even higher currents, and this rvould require that they be made
significantly larger. We suggest that, for Soderberg electrodes, diameters up to about 2.5 metres might be
cJnsidered'with existing technology, and possibly even larger diameters in future as experience grows
around the operation of such elecfiodes. As an indicatiolL equation (4) indicates t}.,al a 2.5 metre electrode
should be abie to carry about 217 kA with a.c. and this current would probably be even higher with d.c.

Limits on the electrical 'resistance' of the burden

By 'resistance' here, we are referring in a generic sense to any one of several possible measures of the
eitent to which the contents of the fumace resist the florv of current when a voltage is applied. This
'resistance' tvrll depend on the interaction between the metallugical processes that occur in the hearth, and
the mode of electrical conduction and dissipation of this power in this same region. The mechanisms will
vary not only lrom one type of feno-alloy process to another, but also from one t)?e of raw material to
another.

It is beyond the scope of this paper to report a detailed study of the conduction mechanisms for each of
these types of proceis. ln summary, however, raising the electrode will raise the 'resistance' in all cases,
but thi; will alio affect the metallurgy. Hence for most existing types of feno-alloy process, with given raw
materials and a given t1,pe of process, there will be an optimum band for the value of the resistance beneath
the electrode. Ii rvill uiually-be desirable to operate towatds the upper end of this band simply to get as
much power in as possible, but it is not possible to increase the resistanc€ indefinitely without adversely
affecting the metallurgical performance.

ln some cases,it may be possible to change the metallurgy to obtain more of an open-alc type of process.
By this means, signifrcantly higher voltages can be achieved. For example, a ferro-alloy like
f#ochromium, which is normally made via a wet process in a submerged-arc fumace, can also be made in
a d.c. arc furnace. In this case, the voltage on the fumace can be made sigrrificantly higher ifthe
process is
operated in an open-arc mode, or in a mixed conduction mode as appears possible for high
power
ferrochromium production.

2002 Electric Fumace Conference Proceedings 431

 CHARACTERISTIC CTIRVES OF THE POWER ST]PPLY

Tbe essential parts ofthe equivalent circuit ofa conventional a.c. power supply are summarised in Fig. 1(b).
ln practice, there is more than one electrode - typically three. The full equivalent circuit will then consist of
a star arrangement with a common hearth, and the source ol electrical power is then a tkee-phase supply
brought in a symmetrical phasor fashion to the three electrodes. However, except for the asymmetrical
interaction effect as discussed further below, the characteristic behaviour of such multi-electrode furnaces is
the same as for the single-phase equivalent circuit shown in Fig. l.

Magnetic inductance behaves as a reactance circuit element in an a.c. circuit, and this involves a voltage
drop when a current flows. Hence as the fumace draws current, so the voltage available to the fumace
decreases. At high resistances and low currents, the voltage drop across the reactance is negligible, and so
the power drssipated in the fumace simply rises in proportion to the current. At the other extreme of low
resistances and high currents, the voltage drop across the reactance domi.nates, and simply limis the current
to a fixed value, so the power falls to zero as the resistance of the furnace drops to zero. Hence the
characteristic curve of this tlpe of power supply is the slightly-skerved, inverted-U-shaped curve that is
sketched in Fig. 2.

When the fi:mace is running, each resistor will give rise to the dissipation of real power Pr, which is the
electrical power that goes into generating the heat in the furnace, and is normally measured in megawatts
(MW). However, the reactance will dissipate only reactive power p, which draws no neft power out of the
mains. This reactive power is normally measured in MVAR to distinguish it from the real power. The total
power is a combination of the real and the reactive power, and is given by the formula a/P'?+Q'z It is .
measured in lvfvA. The power factor, commonly denoted by cos g, is the ratio of the rcal power to this total
power.

As furnace size is scaled up, so the resistances decrease while the reactances increase. These trends can be
exolained as follows:

(a) Resistance - According to the Andreae' and Kelly' formulae, the resistance per phase Rs is related to the
electrode diameter Dg by:

RE: bA|DE (6)

where the constant of proportionality *r remains relatively fxed for a given process. Hence, as the fumace
size increases and the electrode diameter also increases, so the resistances decrease by the reciprocal oftheir
diameter.

(b) Reactance. - We have found that at 50 Hz the individual electrode reactance Xs in a fiunace rvith three
separate transformers and well desigled bus-bars can be related approximately to the electrode diameter D6
by the follo*ing formula:

X6: 0.8+0.2xDs (7)

(For 60 Hz, the reactances will be 1.2 times higher than the value calculated from this formula.)

As a result of these relative trends ofthe resistance and reacta:rce with flrmace size, the reactive power I
rises more steeply than does the real power P. Hence the power factor of the fumace decreases as the
fumace size increases. A low power factor gives rise to a number of operational problems, which is tvhy
larse fumaces tend to be more troublesome.

432 2002 Electric Fumace Conference Proceedings

A direct consequence of having a Q that is comparable to oI greater than P is that the required MVA
larger than the
capacity ofthe tiansformer (and1f tt'e conesponding switchgear and cables) is significantly
MW power ofthe furnace. Figure 5 shows how this tvorsens as the size ofthe fumace increases.

400

300
FeSiMn FeCr

200
=(5 FeCr in high
o arc moqe. ,.
l- -
{00
K''-' *
0
50 100 150

Total MW
Fig. 5. lncrease in transformer MVA capacity with increase in fumace power'
for three different ferro-alloy processes

Area density of power, and gas permeability of the burden

current i,
Ifthe formula for the limiting electrode current 1r from equation (4) is substituted for the electrode
and the formula from equation (6) is substituted for the electrode resistance R5
in equation (1), then the
followinq can be derived:
pa : ?.Re : [55 x DEtsTz.lkA I DEf

: {552 x kn1 x P} (8)

electrode diameter'
This equation (8) shows that the power dissipation is proportional to the square of the
This indicates, inter alia' that the
u.Jtftri. .mpfy proportional to the cross-seciional arei oithe electrode.
dissipation,. will also be
iJ. of g"noution of off-gas, which is roug6ly proportional to the power area of the
pi"pJffu to the cross-seJtional area of *re eteitroae, and hence roughly proportional 10 the
iurien through which the gas must escape. This means that the gas velocities though th: burden
should
this may
i",1uin fuitty""onstant as G fumace is sialed up, which that
should lessen the concem constrarn
the scale up ofthe process, provided a uniform gas distribution can be maintained'

in the burden
However, there is always a problem with hot jets of gas that come up through blowholes
p"AoA"uity. The rate oi generation of oflgas will rise in proportion to the power dissipation, and
ifthis gas
This will impose
iron Ai lr"up"r tfuough ine blowhole, thJheating effect will be proportionately higber.
considerable heat stress on the surounding structues and gas plant

The interaction effect

the actual operation of a c'
The interaction effect is the main consequence of reactance, because it affects
movement affects
iurnu"".. when one electrode in a three-electrode fumace is either raised or lowered, the
in an unsymmetrical
not only the current in that electrode, but also the currents in the other two electrodes

2002 Electric Fumace Conference Proceedings 433

way. This is the interaction effectT. The extent and skewness of the interaction effect increases as the
power factor ofthe fumace decreases.

Soderberg electrodes compound this interaction problem, because they cannot be slipped at rvill to recover
lengtlL unlike graphite or pre-baked electrodes. In particular, if an electrode becomes short for whatever
reason, then the resistance in that phase is likely to be on the high side for a period of time. This will cause
not only the current in that electrode to drop, but will also affect the currents in the other two electrodes.
These other electrodes will get moved to balance their currents, and in time one will grow longer while the
other grou,s shorter. When the original short electrode eventually recovers, the shorter ofthe other trvo may
become the next short electrode, and the sequence will start again. This chronic electrical imbalance will
create a highly skewed power distribution inside the fumace, which may affect the metallurgical behaviour
adversely.

A rough guide as to how the power factor, because of the interaction effect, affects the operability of a
furnace with Soderberg elechodes is as follows:

cos p:0.8 to 1.0: the interaction effect can be observed, but is not a major problem,
:
cos p 0.65 to 0.8: the effect is significant enough to be an occasional problem,
:
cos p 0.5 to 0.65: the effect is likely to be a continual operational problem,
cos p:
<0.5 the effect will be a major operational problem, and the fumace may be almost
unmanageable.

Because the power factor decreases as the size ofthe furnace increases, the interaction effect is probably one
ofthe main constraints on building larger a.c. furnaces.

DIRECT-CURRENT (D,C.) FI]RNACES

Electro-magnetic inductance creates reactance only rvith altemating current. Hence with d.c. power, there
are no constraints from the reactance side. In particular, with a multi-elechode d.c. fumace, each electrode
behaves individually - i.e. there is no interaction effect.

A d.c. porver supply can also bs tailored to provide whatever characteristic curves are required by the
fumace. Thus d.c. fumaces allow much a better conhol of the current, voltage and power being dissipated
beneath an electrode, and, in the case of multi-electrode fumaces, each elechode can be controlled
individually.

Because the voltages on adjacent electrodes in a multi-electrode furnace are similar, and the electrodes do
not interact with each other operationally, tley can in theory be mounted closer together, should the need
arise. It is also worth noting that the flares in a multi-electrode d.c. fumace wrll be directed inwards, which
is the opposite of the direction in an a.c. fumace.

A d.c. fumace requires a high-current connection to the hearth. Recent publications conceming the
operation of open-arc d,c. furnaces have indicated that currents ofup to 180 kA can be conducted from the
bottom ofthe fumace, but these fumaces are constructed for regular headh or anode replacement. Although
the demands on the hearth anodes in submerged-arc mode are much less severe, the questions remain as to
how much current can be removed from the bottom of the furnace, and how long such connections can be
kept functioning. In a very large three-electrode d.c. fumace for feno-alioys, currents of the order of half a
milUon amos wonld need to be conducted out via the hearth connections.

Aa 2002 Electric Furnace Conference Proceedings

^
 DISCUSSION

It would seem from the above analysis that the main factor that constrains the sizes of present a.c.
submerged-arc ferro-alloy fumaces is the inductive reactance caused by the magletic fields around (mainly)
the electrodes in the fumace. The main consequence of inductive reactance would seem to be the
interaction effect. A second consequence is the larger transformer required. Both these consequences
become more severe as the power factor decreases and they therefore become more problematical as the size
of the fumace increases.

The Soderberg electrodes in larger furnaces also seem to be troublesome, but it is not clear whether this is
primarily because of an inherent fragility of the larger electrodes, or because of the interaction effect, or
because of the skin effect. Operation in dc. mode would alleviate the interaction effect and the skin effect,
but not the inherent electrode fragility.

With most existing submerged-arc prqcesses, there is little that can be done about significantly increasing
the resistance ofthe burden. However, as in t.he case of ferrochromium, it may be possible to change to a
fully open-arc or a mked mode of operation, thereby increasing the resistance (and therefore the voltage)
significantly. ln the case of ferromanganese alloys, the high vapour pressure of manganese at the normal
fumace temperatures will limit the degree of arc power without incurring significant metal losses. WitI
silicon, and ferrosilicon, the reaction b€tween silicon monoxide and carbon requires the presence of a solid
permeable burden above the reaction zone around the arc. Sigaificantly longer a.c. arcs therefore seem to be
undesirable in the silicon process. On the other hand, with ferronickel, the normal mode of operation is
semlopen, and thus very high voltages are tle norm. To *hat degee the use ofd.c. power for open-arc and
semi-open-arc operation will benefit these other processes has possibly not been fully evaluated.

With the corect pre-treatment of the feed materials, the permeability of the burden may not be such a
constraint at higher power levels, but the fumace infrastructure will need to be able to handle the occasional
hot "blow''.

it is evidentfrom tlre above arguments that d.c. operation could provide a way to increase the size of
fumaces sigmficantly beyond what is currently attainable with a.c. furnaces. Furthermore, it is not
inconceivable that, with the use of multiple-el€ctrode d.c. technology, ferro-alloy fumaces of well above
100 MW could be built and successfully operated.

The use of graphite or pre-baked electrodes may also help to overcome some of the problems inherent in
large Soderberg electrodes, as they do not require the same level of care in their operation, and in particular
lhey do not ha\ e to be slipped evenly and regularly.

From the vie*'point of the electricity supply grid, existing a.c. submerged-arc fumaces do not impose any
major problems, apart from low porver factors, *'hich can in any case be corrected by capacitors should the
need arise. Flicker is not normally a problem with submerged-arc furnaces, udike their open-arc relatives.
Furthermore, the technology for delivering a.c. power to a fumace is relatively basic, well established, and
trouble free. The d.c. fumace, ho1ever, is not so simple. The tlpes of problems created dep€nd on the tlpe
of porver supply. Thlristor bridge supplies generate severe harmonics in the waveform of the current drawn
flom the mains, and because ofthis, considerable effort has been required in the development of active and
passive compensation circuits. The best solution for minimising the disturbance to the mains is the use ofa
ihopper supply, also called a pulse-width modulation (PWM) supply, which often uses insulated-gate
bipolar transistors (IGBTs). However, tlis tlpe of power supply is relatively bulky, complex, expensive,
and its reliability has yet to be proven.

It is also not knox.n to what extent problems would be created by the electrical load of a very large ferro-
alloy fumace on the electricity supply gnd, e.g. by simply s*'itching on or offCould there be any fansient
instability problems, and would the supply authority tolerate such loads?

2002 Electric Fumace Conference Proceedings 435

The production from very large furnaces relative to the total world production may become an issue if the
fumaces are built significantly larger. ln the case of high-carbon ferrochromium for example, the existing
largest fumaces can each produce about 2o/o of the total world production. For larger furnaces still, the
percentage would be proporlionately higher.

At the peak of the characteristic curve in Fig. 2, the power factor is 0.7, while to the left of the peak the
power factor is higlrer and to the right it is lower. Hence operation to the right of the peak of the
tharacteristic cuwe is associated with large fumaces and with operational problems. Fig. 6 shows an
approximate line of constant power factor, superimposed on the trend lines fiom Fig. 3. This line
conesponds to a cos g of about 0.7 at 50 tIz- To the left of this line, the power factor is higher than 0.7,
while to the right it is lower. That is, this line forms the approximate boundary between the problematic
operating region to the right, and the more stable region to the left. This figure shows that, in general, the
trinds of all the processes are aiming towards the more problematic region of operation. However, the
ferrosilicomanganese process is already well into this region, while the ferrochromium process still has
some way to go.

Fecr in high
=
=
arc inode
o
o
E
F

120

Electrode current, kA
Fig. 6. A line ofconstant power factor, superimposed on the trend lines ofFig. 3, to indicate the
approximate boundary ofthe problematic region ofoperation in a.c. fumaces

CONCLUSIONS

Conventional submerged-arc fumaces that use a.c. mains power directly would seem to have reached or are
close to a limit in size. This limit is caused primarily by the low resistance compared to the reactance in the
fumace, which leads to low power factors, and this incurs operational problems that are severe enough to
discourage the installation of large fumaces. There may also be a separate problem rvith self-baking
Soderberg electrodes as they get larger, which also discourages the installation of larger fumaces. Pre-
baked carbon or graphite electrodes *'ould overcome most ofthe problems with Soderberg electrodes.

In recant ).ears, power supplies have become available that can deliver
d.c. power at very high crments
From the point of view of the fumace operation, d.c. obviates the problems of reactance. There is also a
possibilitythat d.c. may lessen the problems with large Soderberg electrodes, but this remains to be proven

436 2002 Electric Fumace Conference Proceedings

While d.c. operation would seem to have relaxed some ofthe limits on conventional fumaces, there are still
a number ofquestions that are unanswered, such as:
. To what extent does d.c. operation affect the mechanisms and alter the metallurgy of existing t)?es
of feno-alloy production processes?
. AJe there any restrictions on the currents that can be carried via a comection to the hearth?
. How reliable is a very large d.c. power supply likely to be?
. Very large furnaces would constitute a significant load on the electricity grid. Would this present
any problems for the stability and control of the grid? Would power supply authorities tolerate such
large unit loads?
o Are there any extemal constraints on the sizes of feno-alloy fumaces, such as marketing and
distribution of the products?
. Hot blowholes though the burden in the fumace may become an issue with sigtificantly larger
fumaces, necessitating greater attention to the control ofthe size characteristics of feed materials.

Summary
This paper examines the tecbnical issues thal cunently constrain large-sized ferro-alloy fumaces. The
possiUitity of further scale-up is also studied, and each ofthe various constraints is investigated with a view
io horv ii might be eased oi even circumvented. It would appear that d.c. power may offer some critical
advantages ii this regard, and may even help to relieve some of the problems that existing large fumaces
display.

Ackn owledgemenls
This paper is published by permission of Mintek

References
L N A Barcza, "Recent Technical Developments and Future Trends in the Southem African Feno Alloy
Industry', Mebl Bulletin Ferro Allov Conference, February 2002, Sun City, South Africa.
2. F. V. Andreae, "Design and Control ofFerro-Alloy Fumaces", AIEE Transactions. Vol. 69, i950, pp. 557-562-
3. W. M. Kelty, "Design aad Consfiuction of the Submerged Arc Furnace", Carbol and Graphite News. Vol. 5' No
1, AprllMay 1958.
4. J. Westlv. 'Dimension des fours de reduction pour Fe-Si et d'autr€s Ferro-alliages", Journal
du Four Electrique.
No. 1, January 1979,PP.14-19.
5. A. L. Moolman, Ir,l.-S. Rerurie, P. Brereton-Stiles, "Advanced Power Control Strategy for Submerged-Arc
Fumaces"
Canada The Ferroalloys Assoc., Washingloo, DC, USA June 2001, pp 103-i07'
6. .q.. dewaal, I. J. Ba*er, tr,t. S. nennii, J. Kopper, B. S. Groenevel4 "Electrical -factors
affecting the
of the
* Intemational
6'Intemational
economic optimisation of submerged-arc fumails", INFACON - 6 Proceedi{rss
'Conference.
Fenoallols Cape Town. Vol. 1. Johannesburg, SAIMIT{, 1992,pp247-252
7 I J Bu.ker. A B St"*art, "ktdoctiv" reactance, and the operation of large submerged-arc fumaces"' I-L
Afr. Inst. Min & Metall., Vol. 80, No 3, March 1980, pp. 123-128'

2002 Electric Furnace Conference Proceedings 437

 60th
ELECTRIC FURNACE
CONFERENCE PROCEEDINGS

November l0 - 13,2002
San Antonio, Texas

Sponsored by the
Electric Fumace Division
ofthe
Iron & Steel Society

David L. Kanagy
Publisher

Margaret A. Baker
Editoa Book Publishing
 The Iron & Steel Society
is not responsible for statements or opinions
expressed in this publication.

60th Electric Fumace Conference Proceedings, Vol. 60

These proceedings are also available in CD-ROM format.

Copyright @ 2002
Iron & Steel Society
All rights reserved
Printed in the U.S.A.

ISBN: 1-886362-63-7
ISSN:0096-0128

ISS Publications...Ktowledge fulfillment with a commitment to the

quality, ifiegrtty and timeliness of information transfen

Iron & Steel Society

186 Thom Hill Road
Warrendale, PA 15086-7528
Phone: Q24)776-1535
Fax:. (724)776-0430
E-mail: [email protected]
Web site: www.iss.org

Publication supported by the Seeley W. Mudd Memorial Fund of AIIT4E