http://dx.doi.org/10.

17159/2411-9717/2018/v118n6a11

Basic parameters in the operation and

design of submerged arc furnaces, with
particular reference to production of
high-silicon alloys
by T.E. Magnussen*

electrode positions, and at the same time

$() maintaining a constant furnace load.
A review of the basic parameters in the operation of submerged arc furnaces The operating resistance depends on the
is given. The importance of controlling the electrical operating parameters in required quality of the product, the raw
accordance with the principles presented by Jens Westly in the 1970s is
material properties, the composition of the raw
highlighted.
Basically, the specific energy consumption for production of high-silicon
material charge mix, and optimum utilization
alloys depends on the silicon recovery and the furnace efficiency. Silicon of electrical energy.
recovery depends mainly on the raw material quality, but also on furnace For a given process and charge mix
design. The furnace efficiency depends on the furnace operating resistance composition, it is a matter of experience to find
and the loss resistance, which are again related to the furnace design. the optimum operating resistance.
A case from practical furnace operation is presented, showing how the In the early days of the ferroalloy industry,
furnace dimensioning and design affect the furnace efficiency and hence the
the fundamental formula in furnace design
specific energy consumption. Possibilities for further improvement of furnace
efficiency and reduction of the loss resistance are discussed. was the Andreae formula, introducing the k-
factor (Andreae, cited in Westly, 1974):
&('
submerged arc furnaces, high-silicon ferroalloys, electrical operating  [2]
parameters, specific energy consumption, furnace efficiency factor, loss
where R is the operating resistance in m
, D
resistance.
is the electrode diameter in metres, and k the
peripheric resistance (k-factor) in m
-m.
/$"'(!#")($ According to Andreae, the value of the k-
factor varies with the process, the raw
It is the author’s impression that chemical
materials, and the specific load of the
reactions and equilibria have normally been
electrodes.
the main focus in the evaluation and studies of
Later, adjustments were proposed to the
process performance of high-silicon alloys
Andreae formula by Kelly (Kelly, cited in
production. Electrical operating parameters did
Westly, 1974), Persson (Persson, cited in
not always receive the same degree of
Westly, 1974), and others, suggesting how the
attention from metallurgists.
value of the k-factor depended on the power
In the author’s experience, it is very
density in the electrodes.
important to also pay attention to electrical
Jens Westly studied operating data from a
parameters in furnace operation.
large number of electric reduction furnaces,
and published the result of his work on several
&'%")$*'&)"%$#&*%$*"&*-&"
occasions during the 1970s (Westly, 1974,
%#"(' 1975, 1979). These references are highly
For a three-phase furnace, the operating recommended.
resistance R in m
is defined as follows: Westly suggested the following correction
[1] of the Andreae formula:
[3]
where P is the furnace power in kW, measured
on the primary side of the furnace where k is the k-factor in m
-m, and i the
transformers, and I is the electrode current in electrode current density in A/cm2.
kA.
The operating resistance is selected by the
furnace operator within certain limits,
determined by the furnace mix composition, to * Vatvedt Technology AS, Norway.
© The Southern African Institute of Mining and
obtain a satisfactory electrode penetration and
Metallurgy, 2018. ISSN 2225-6253. This paper
good tapping conditions. was first presented at the INFACON XV:
The normal way of controlling the International Ferro-Alloys Congress, 25–28
operating resistance is to operate the furnace February 2018, Century City Conference Centre and
on constant electrode current by adjusting the Hotel, Cape Town, South Africa


   

 

VOLUME 118  
 631
Basic parameters in the operation and design of submerged arc furnaces
By combining Equations [2] and [3], Westly (1974) If R is too high compared to RC, the proportion of the total
deducted the following correlation: energy input consumed in the melting of raw materials will
be too high, and too little will be available for reduction.
[4]
This phenomenon is well-known from production of
where I is the electrode current in kA, and P the furnace high-silicon alloys. If a furnace is operated on too high an
power in MW. operating resistance compared to the resistance of the charge
Based on the above, Westly introduced a new factor, the mix, too much of the energy input will be used for melting of
Westly factor (C3) which is defined as: raw materials compared to the energy available for reduction
of silicon. This will result in an excess of slag in the furnace
[5]
and insufficient reduction of silicon, as well as impurities
For a given process and a given raw material mix such as aluminium and calcium. This situation may be
composition, the value of the Westly factor (C3) is basically interpreted as a problem with the carbon balance, whereas
constant and independent of the furnace load. Reference is the real cause of the problem is that the operating resistance
made to the furnace load diagram given in Figure 1. is too high compared to the resistance of the charge mix.
The Westly factor has been found to be extremely useful Typical signs are slag formation and low metal
as a guideline for furnace operation and dimensioning, temperatures, causing tapping problems. Further, the
particularly in relation to upscaling of furnaces. electrical instruments will tend to show little response to
raising or lowering of the electrodes, indicating that the
&*')$#)&*( *&%"*)"')!")($ current does not pass from the electrode tips, but from the
sides of the electrodes, causing abnormal electrode wear.
Westly also (1974) introduced the principle of heat
To conclude, the operating resistance cannot be chosen at
distribution.
random, but in due consideration of the resistivity of the
Equation [1] rearranged gives the formula for the furnace
charge mix.
load P (kW):
[6] &*)$) )#%$#&*( *"&* !'$%#&*'&%#"%$#&
where R is the operating resistance in m
and I the electrode For a three-phase furnace, the furnace reactance X in m
is
current in kA. defined as follows:
With reference to the above formula, furnace operators X [8]
will normally tend to operate on the highest possible
operating resistance in order to obtain the highest possible where Q is the reactive furnace power in kVAr, measured on
load on their furnace. However, there is a practical limit for the primary side of the furnace transformers, excluding the
the operating resistance that can be used, determined by the load of capacitor banks, and I is the electrode current in kA.
heat distribution factor C: Whereas the furnace load is depending on the operating
resistance and the electrode current, it is independent of the
[7]
furnace reactance.
where PC is the energy generation in the furnace charge, P is The reactance must be taken into account when designing
the total energy generation, R is the operating resistance, and the transformers for a new furnace. The furnace reactance
RC the resistance of the furnace charge. affects the required transformer capacity as well as the
C may take values between 0 and 1. As RC is given by the secondary voltage range for the transformers. It also affects
raw material mix composition, C is controlled by the choice of the size of the capacitor banks for correction of the power
the furnace operating resistance R. factor on the electrical power supply grid.

)!'&*(%*)%'%* ('*%$*&&#"')#*'&!#")($* !'$%#&



632  
 VOLUME 118    

 


Basic parameters in the operation and design of submerged arc furnaces

)!'&*)#%*(#%")($*( *#!''&$"*%"* ('*) &'&$"* !'$%#&*#'%"&'*#($)")($

Somewhat simplified, we can say that the furnace &#"')#%*%'%&"&'* ('*&%#*&&#"'(&

reactance is proportional to the area surrounded by the
The resistance and reactance values calculated as described
current circuit.
above give the average values for the electrodes in a three-
The furnace reactance increases with increasing electrode
phase furnace, based on the active and reactive power
spacing, and decreases with increasing electrode diameter. A
measured on the primary side of the furnace transformers.
high electrode holder position and low electrode tip to hearth
It is also interesting to establish electrical measurements
position, with current paths located at a low level in the
for each electrode in order to monitor and control the
furnace crucible, will give a high furnace reactance, and vice
conditions on each phase.
versa.
These measurements have to be based on a secondary
Based on experience, it is worthwhile to register and
voltage measurement between each phase and the furnace
study the value and the trends of the furnace reactance in
hearth. For this purpose, we measure the voltage between
daily operation, since it gives an indication of factors such as
knife contacts on the electrode casing and an artificial zero
the electrode length and electrode penetration. In other
point underneath the furnace.
words, it gives interesting information regarding the
operating conditions of the furnace. Based on the secondary voltage and the secondary
Figure 2 indicates the typical locations of the current current measurement, it is possible to measure active power,
paths for different furnace crater conditions. reactive power, resistance, and reactance for each electrode.
The challenge with these measurements is the strong
!'$%#&*(&'* %#"(' electromagnetic field surrounding the furnace, which causes
induction in the cables for voltage measurement, thereby
A consequence of the Westly formula is that the operating
influencing the accuracy of the electrical data. Even though
resistance R decreases with increasing furnace size.
we try to reduce the influence of the induction by
Referring to the expression for the furnace power factor:
symmetrical arrangement of the cabling around the furnace
C x [9] and adjustable resistances in the measurement circuits, there
will still be some inaccuracy in the voltage measurements and
where Z is the furnace impedance in m
,
the electrical data deduced from them.
a natural consequence of increasing furnace size is that
the furnace power factor decreases. The power factor of large
furnaces today is normally less than 0.707. At the time, some %#"('*% &#")$*"&*&#) )#*(&'*#($!")($* ('
people thought this may be a problem for the furnace )0))#($*%(
operation; but normally, it is not.
It is well known that the specific power consumption
However, with decreasing power factor, it is becoming
(kWh/mt Si) depends on the silicon recovery.
more difficult to balance the power load and the charge mix
Above all, the silicon recovery is dependent on the raw
consumption between the electrodes in the furnace. If the
material quality and the charge mix composition. The use of
power factor is very low, towards 0.5 and lower, this balance
reductants with high reactivity versus SiO gas will increase
becomes quite critical, since the electrode to hearth resistance
the silicon recovery. The use of reductants of smaller size will
is fairly close to the short circuit condition
increase the available surface area of the reductants for
Today, most power supply companies require installation
exposure to SiO gas, and will therefore be beneficial for
of equipment for power factor compensation in order to avoid
too low a power factor on the power grid. Capacitor banks are increasing the silicon recovery.
installed for this purpose. The most common, simple and Furthermore, the silicon recovery is also affected by the
sturdy arrangement is to install capacitor banks in shunt furnace design.
connection on the primary side of the furnace transformers. The following features of our furnace design are
However, if the furnace power factor is very low, it is implemented to improve the silicon recovery:
normally preferable to install the capacitor banks in series. In  Design of day bins and furnace bins to avoid
this case, the problems of unbalance of the load and charge segregation
mix consumption between the electrodes as described above  Rotating charging tubes
can be avoided.  Rotating furnace body.


   

 

VOLUME 118  
 633
Basic parameters in the operation and design of submerged arc furnaces
In particular, the use of rotating charging tubes has furnace operator as mentioned above, RL is determined by
turned out to be very beneficial in order to improve the the furnace design, physical dimensions, and electrical
silicon recovery. conditions in the furnace crucible.
However, the specific power consumption is also The furnace operating resistance R depends on the charge
dependent on the furnace efficiency factor. mix composition and the raw material quality. The main
The furnace efficiency factor indicates how efficiently the factors that cause an increase in the operating resistance R
supplied electrical energy is being utilized in the furnace: are the following:
[10]  Smaller sizing of reductants
 Increased proportion of coal in charge mix
where  Use of wood chips.
 is the furnace efficiency factor The above measures should be used and are being used
P is the furnace power in MW
by furnace operators today to increase the furnace efficiency
PL is the power in MW lost by ohmic resistance in secondary
and hence reduce the specific energy consumption. These
conductors and electrodes, as well as by induction loss and
measures come in addition to the effect of using more
other electrical losses.
reactive reductants in order to improve the silicon recovery as
The specific power consumption for production of high-
mentioned above.
silicon alloys can be expressed by the following formula:
The furnace loss resistance RL depends on furnace design
[11] and dimensioning. The following factors contribute to
decreasing loss resistance:
where  is the furnace efficiency factor,  is the silicon
recovery ( = 1 corresponds to 100% silicon recovery), and  Materials with high electrical conductivity in furnace
K1 and K2 are constants. components
With reference to the data given by Schei, Tuset, and  Optimized cross-sectional area of current conductors
Tveit (1998), the formula for the specific power consumption  Nonmagnetic materials to reduce eddy currents.
can be written as follows: We take all these factors into account as part of our
[12] furnace design, with the aim to achieve the lowest possible
loss resistance.
Figure 3 shows the specific power consumption as a In addition, the loss resistance can be influenced by the
function of the silicon recovery for different values of the conditions in the smelting crater, as shown in the example
furnace efficiency factor. below.
The specific power consumption decreases with
increasing furnace efficiency and increasing silicon recovery. ,%&* '(*'%#")#%* !'$%#&*(&'%")($
In order to reduce the specific power consumption, we need to Equation [12] is very useful for analysing furnace
affect these two factors. performance, as demonstrated by the following example.
The furnace efficiency factor can also be expressed by the The author once had the opportunity to study two silicon
following formula: metal furnaces in the same plant and of the same capacity,
operating at about the same furnace load, using the same raw
[13]
materials and charge mix composition.
where R is the total resistance or operating resistance, as The strange thing about these furnaces was that, in
referred to above, and RL is the loss of resistance in simplified terms, they had about the same silicon recovery,
secondary conductors, electrodes, and by induction. but not the same specific power consumption.
The furnace efficiency increases with increasing R and Table I lists the data for the two furnaces, A and B, for a
with decreasing RL. Whereas R is basically selected by the given reference period. Using the above formulae, the furnace

)!'&* &#) )#*(&'*#($!")($* ('*'(!#")($*( *))#($*&"%*



634  
 VOLUME 118    

 


Basic parameters in the operation and design of submerged arc furnaces

Table I Table II
&'%")$*%"%* ('*))#($*&"%* !'$%#&*%&*($ &'%")$*%"%* ('* !'$%#&**
*%&*($
)!)*&"%*'(!#")($ )!)*'(!#")($*%)** &''())#($
!'$%#&
 Electrode diameter, mm 1 200
Electrode spacing centre to centre, m 2.70
Inner crucible diameter, m 6.80
Electrode diameter, mm 1 143 1 205
Furnace load, MW 18.04
Approximate electrode spacing centre to centre, m 2.80 2.74
Operating resistance, mΩ 1.244
Approximate inner furnace crucible diameter, m 6.60 6.60
Westly factor (C3) 10.11
Furnace load, MW 19.50 19.90
Furnace reactance, mΩ 0.983
Operating resistance, mΩ 1.333 1.309
Specific power consumption, kWh/t 7 997
Westly factor (C3) 9.64 9.70
Silicon recovery, % 93.60
Furnace reactance, mΩ 1.310 1.273
Furnace efficiency factor 0.857
Specific power consumption, kWh/t 12 371 11 795
Loss resistance, mΩ 0.178
Silicon recovery, % 82.17 84.55
Furnace efficiency factor 0.826 0.844
Loss resistance, mΩ 0.232 0.205
Short circuit resistance, mΩ 0.313 0.278
Short circuit reactance, mΩ 1.477 1.202 ($&!&$#&*( *)$#'&%)$* !'$%#&*)1&
As mentioned above, the operating resistance decreases with
increasing furnace size. Normally, the loss resistance also
decreases with increasing furnace size, due to the increased
efficiency factor  and the loss resistance RL were calculated. cross-sectional area of the current conductors. Therefore,
In addition, a measurement of the short circuit resistance and generally, increasing the furnace size has no adverse effect
the short circuit reactance was carried out by melting down on the furnace efficiency.
the furnace and taking the electrical readings with the However, large furnaces are more sensitive to variations
electrodes resting on the furnace hearth. in raw material quality. A consistent quality of raw materials
As shown in the table, the furnace efficiency factor is and electrode paste is very important in order to achieve
higher and the loss resistance lower for furnace B than for stable operating conditions and good efficiency for large
furnace A. In addition, the silicon recovery is somewhat furnaces. The consequence of operational disturbances and
higher. furnace shutdowns are normally more severe for a large
It is thought that the reason for the higher furnace furnace than for smaller ones.
efficiency and lower loss resistance for furnace B compared to Generally, the investment cost per megawatt and the
furnace A is the larger electrode diameter and smaller manpower cost per ton of product is lower for large furnaces.
electrode spacing. It seems reasonable that the larger In our opinion, this is the main reason why large furnaces are
electrode diameter on furnace B is at least part of the reason preferable to smaller furnaces for new projects of a given
for the difference in loss resistance. overall production capacity in an established industrial
Moreover, the electrodes had an inclination with the environment.
electrode tips pointing towards the periphery of the furnace
crucible, probably caused by an inactive zone with a solid !'"&'*&+&(&$"*)$* !'$%#&*&)$
cone in the centre of the furnace. This phenomenon was more Vatvedt has recently been working on improvements in
predominant on furnace A than on furnace B. Given this design of the electrode column, including the development of
factor, there is reason to believe that part of the current in the contact clamps and pressure rings of our own manufacture.
furnace crucible was passing from the outside of the Our company supplies contact clamps and pressure ring
electrodes near the electrode tips, out to the carbon sidewall segments made of pure electrolytic copper. The water-cooled
lining and along the periphery of the furnace crucible over to shields are made of stainless steel.
the neighbouring electrodes. The wear pattern of the In particular, we are now looking at the design of the area
electrodes also supported this theory, with higher wear on the of the electrode column above the top of the contact clamps.
electrode surface facing the carbon sidewall lining. Figure 4 illustrates the conventional design of this area.
The short circuit measurements also support the above A well-known problem in operation of furnaces of this
theory. The calculated short circuit resistance was about 13% design using Søderberg electrodes, particularly for large
higher on furnace A than on furnace B, and the calculated furnaces, is that there is an abundance of thermal energy
short circuit reactance was about 23% higher. Furthermore, developed in the area above the contact clamps, caused by
this supports the above comments with respect to how the induction heating due to the strong electromagnetic fields.
location of current paths in the furnace crater has an This surplus heat causes melting and heating of electrode
influence on the furnace reactance. paste; sometimes resulting in overheating of the liquid paste,
As a consequence of the above findings, the electrode paste segregation and high liquid paste levels. This is
spacing on furnace A was reduced, resulting in a substantial resulting in poor quality of the baked electrode, causing
improvement in the furnace performance. disturbances in the electrode operation and increased risk of
Table II lists the data for the furnace supplied by Vatvedt hard as well as soft electrode breakages (Skjeldestad et al.,
to ChEMK in Chelyabinsk, Russia in 2014, for the production 2015).
of ferrosilicon. Compared to the data for the silicon metal When considering this problem in more detail,
furnaces above, the furnace efficiency is higher and the loss calculations have confirmed that with the conventional
resistance is lower. design of this area, the available heat generation by


   

 

VOLUME 118  
 635
Basic parameters in the operation and design of submerged arc furnaces

)!'&*)#%*&)$*( *(&'*%'"*( *&&#"'(&*#(!$

induction heating is much larger than what is required for Above all, the silicon recovery is dependent on the raw
melting and heating of the electrode paste before baking. In material quality and the charge mix composition. The silicon
other words, this is a waste of energy, which is adversely recovery is also affected by the furnace design.
affecting the furnace efficiency. The furnace efficiency factor depends on the operating
Based on some initial studies, it is thought that by resistance of the furnace, which again depends on the
making some design modifications in the area above the top resistivity of the furnace charge mix. Smaller sizing of the
of the contact clamps, it would be possible to achieve a reductants, an increased proportion of coal, and use of wood
considerable reduction of the induction heating in this area. chips will contribute to increased resistivity of the charge
This reduction of energy loss will have a corresponding effect mix, which will make it possible to operate with an increased
on reduction of the specific energy consumption per ton of operating resistance.
alloy. Further, the furnace efficiency factor is depending on
In addition, the problem of overheating and segregation furnace design and dimensioning, which is affecting the
of electrode paste with its adverse effects on electrode quality electrical losses from the furnace. The importance of proper
and operation, which is the case on many furnaces today, is design and material selection for furnace components are
likely to be eliminated. Therefore, it will be a win-win highlighted. The significance of the furnace dimensioning is
situation. illustrated by an example from practical furnace operation.
Even with this reduction of the induction heating above Possibilities for improvements in furnace design, and the
the top of the contact clamps, calculations have shown that electrode column design in particular, were highlighted with
there will still be enough induction heating available for the aim of reducing electrical losses and hence improving
melting and heating of the electrode paste. furnace efficiency even further.

.& &'&$#&
($#!)($ SCHEI, A., TUSET, J.K., and TVEIT, H. 1998. Production of high silicon alloys.
It is very important to pay attention to electrical parameters in Tapir Forlag, Trondheim, Norway
SKJELDESTAD, A., TANGSTAD, M., LINDSTAD, L., and LARSEN, B. 2015. Temperature
furnace operation. The studies conducted and the formulae profiles in Søderberg electrodes. Proceedings of INFACON XIV, the
developed by Westly in the 1970s are very useful as Fourteenth International Ferroalloys Congress, Kiev, Ukraine, 31 May–4
guidelines for furnace operation as well as for furnace June 2015. pp.327–337. http://www.pyrometallurgy.co.za/
InfaconXIV/327-Skjeldestad.pdf
dimensioning, particularly in relation to upscaling of WESTLY, J. 1974. Resistance and heat distribution in a submerged-arc furnace.
furnaces. Proceedings of INFACON I, the First International Ferroalloys Congress,
The specific power consumption (kWh/mt Si) for the Johannesburg, South Africa, 22–26 April 1974. pp. 121–127.
WESTLY, J. 1975. Critical parameters in design and operation of the submerged-
production of high-silicon alloys depends on the silicon arc furnaces. Proceedings of the AIME Electric Furnace Conference,
recovery as well as on the electrical efficiency of the furnace, Houston, TX, 9–12 December 1975. pp. 47-53. http://library.aimehq.org/
expressed as the furnace efficiency factor. The specific power library/books/Electric%20Furnace%201975/Electric%20Furnace%201975
%20-%20019.pdf
consumption decreases with increasing silicon recovery and WESTLY, J. 1979. Dimensions des fours de reduction pour Fe-Si et d’autres
increasing furnace efficiency. Ferro-alliages, Journal de Four Electrique, Janvier 1979. 


636  
 VOLUME 118