furnace uses about 12 per cent less power per ton than a 2.5 to 3 ton furnace of the same type.

In refining molten steel in the electric furnace, the power consumption has varied from 100 to 300 kilowatt-hours per ton. Rail steel of this class takes about 100 kilowatt-hours per ton when one slag only is necessary and about 200 kilowatt-hours per ton with two slags in a 15-ton furnace. Tool steel is made from molten metal in a 2 to 3 ton furnace with a power consumption of 150 to 300 kilowatt-hours per ton.

POWER FACTOR.

The power factor is the ratio of the number of effective watts in an alternating-current circuit to the number of volts multiplied by the number of amperes. Any self-induction in the circuit causes a difference in phase between the electromotive force or voltage and the current or amperage, and the current will thus lead or lag behind the voltage; that is, the two do not reach their maxima at the same moment. An amount of energy equal to the number of amperes multiplied by the number of volts is surging through the circuit, but, as the electromotive force and current do not reach their maxima together, the product does not represent the effective energy that is usefully consumed at the receiving end.

The generators have to supply the total wattage and the line has to conduct it. Hence to offset a low power factor large generators and thick conductors must be used. For example, a power factor of 0.5, or 50 per cent, means that conductors of double size must be used to get the same number of effective watts as with a power factor of 1, and it also means that generators of double capacity must be used to produce the same useful output. In such a case there would still be objection raised by the power company, for the voltage regulation of the line and generators would be poor, and the low power factor would act as a brake on the circuit, throwing back on it all power not effectively used and causing the generating machinery to heat. If the company using a furnace with a low power factor generates its own power, it must consider the increased capital investment on account of larger generating capacity.

If the ordinary loads of a central power station are considered, an operating power factor above 0.95 will be obtained only when practically all of the load consists of synchronous motors or rotary converters that may be operated with a power factor of about 1. The power factor for a plant having a large proportion of induction motors is about 0.70. Hence in operating several electric furnaces the necessity of keeping the power factor of the single unit as much above 0.80 as possible is evident or the power factor of the whole plant may drop below 0.70.

Without discussing the causes in furnace design that influence the power factor, it may be said that there are three ways of reducing the trouble. One is to reduce the frequency of the alternating current, but, although this should be considered when a new plant is to be erected, in an established works such a change means new electrical machinery of special design and higher cost. A more practical way is to increase the resistance of the furnace, which results in a higher working voltage and a higher power factor. The third method is to decrease the quantity of iron surrounding the brickwork of the furnace as much as possible, as this iron causes the formation of lines of force that tend to lower the power factor. Tie-rods and buck staves may be made of nonmagnetic special steel, or a strip of copper may be inserted to break the continuity of the iron surrounding the circuit.

The effect of frequency has been mentioned. It is for this reason that most electric steel furnaces are operated with a current having a frequency of 25 cycles rather than with the 60-cycle current ordinarily used in commercial practice. Up to the time of the design of the combination induction furnace the great drawback to the induction furnace as originally built was the necessity of using a current with a frequency as low as 5 to 15 cycles in order to obtain a power factor over 0.5. With the combination induction furnace a power factor of 0.6 can be obtained with a frequency of 25 cycles.

The power factor of a furnace decreases if the size is increased and the voltage and frequency remain the same. For instance, the power factor of a 25-ton single-phase Héroult furnace varies from 0.85 to 0.95 with a current frequency of 25 cycles and a voltage of 110 volts, whereas in a 15-ton three-phase furnace the power factor is 0.70 to 0.80 with a current frequency of 25 cycles and a voltage of 90 volts.

The power factor of the combination induction furnace varies from 0.60 in an 8 to 10 ton furnace, single-phase, frequency 25 cycles, to 0.70 in the 3-ton furnace, three-phase, frequency either 25 or 50 cycles. The power factor of the 1.5-ton induction furnace is about 0.5, using a current with a frequency of 15 cycles. The power factor of induction furnaces, owing to the large proportion of iron used, decreases rapidly with increase in size, so that it would seem that this might prevent an increase in size over 8 to 12 tons.

In general the arc furnace with the conducting hearth tends to have a slightly lower power factor than the arc furnace with the upper electrodes in series and a nonconducting hearth. The power factor of all designs of commercial electric steel furnaces of the arc type is above 0.70 for the sizes over 5 tons, and above 0.80 for the smaller sizes.

ELECTRODES.

With improved methods of making amorphous-carbon electrodes the electrode consumption of electric steel furnaces has been considerably reduced. The manufacture of electrodes with threads for continuous feeding has reduced the stumpage loss, which was formerly as much as 50 per cent in some cases, and the electrodes are now much more uniform in character. Threaded amorphous carbon electrodes of foreign manufacture sell for about 2 to 3 cents per pound f. o. b. at the maker's plant. Domestic electrodes sell at about the same price but are not made threaded for continuous feeding. The use of graphite electrodes for steel furnaces is not economical because of the high cost per pound and because the consumption per ton of steel produced is almost as large as that of amorphous-carbon electrodes. The manufacturers of graphite electrodes claim that the gain in electrical conductivity offsets the higher cost, although the losses by heat conductivity would be greater unless the graphite electrodes were carefully designed.

For melting and refining cold scrap the consumption of amorphouscarbon electrodes varies from 15 to 50 pounds per ton. The consumption seems to be considerably higher in furnaces of the nonconductinghearth type with two electrodes than in the conducting-hearth type with one electrode. The difference is probably due to the additional electrode, for in conducting-hearth furnaces of larger size having two to four electrodes the consumption per ton is greater than in the single-electrode furnace. In furnaces used for refining molten steel the wear on the electrodes is not as great and the consumption is much lower, being 6 to 12 pounds per ton.

THE LINING.

The rate of wear of the lining varies so much that it is almost impossible to state general figures. The furnace bottom will last 200 to 400 heats on cold charges and indefinitely on hot metal. The sides are repaired about every 100 heats but are not completely torn out. In most furnaces the hearth has a dolomite bottom and magnesite brick on the sides. The roof is generally silica brick, and, in an arc furnace, lasts 40 to 120 heats. In the arc furnaces operating under the best conditions the roof lasts about 70 heats. In furnaces with magnesite hearth, sides, and roof the lining cost per ton of metal is high. In some of the more recent furnaces the lining on the sides is about half as thick as that formerly used. The thinner lining increases the capacity and does not seem to affect the conduction losses enough to increase decidedly the power cost per ton of output. The hearth lining of an induction furnace lasts about 55 heats and the roof an indefinite period.

ESSENTIAL FEATURES OF REFINING IN AN ELECTRIC FURNACE.

a

The general process of refining steel in the electric furnace, as has been stated, comprises oxidation, followed by reduction or deoxidation. Owing to the nature of the heat supply, it is possible to have the atmosphere oxidizing, reducing, or neutral. This is not possible in the open-hearth furnace owing to the gases introduced for heating. The function of slags in refining steel in electric furnaces is somewhat different from that in refining in open-hearth furnaces. First, all oxygen introduced into the electric furnace must be in the solid form, whereas the open hearth has an unlimited supply in the air admitted. Oxidation can be conducted in an ordinary openhearth furnace as in an electric furnace, except that the higher temperature of the latter hastens the melt. Deoxidation in which carbon is the chief agent can be performed only in the electric furnace.

During the oxidizing period sulphur is removed only to a small degree in the open-hearth furnace but much more efficiently in the electric furnace, especially when manganese ore is used. It is likely that part of the sulphur forms sulphur dioxide and passes off with the gases. The phosphorus is eliminated in the oxidizing period, just as in the open-hearth furnace.

Sulphur is removed in four ways: (1) As calcium sulphide, formed by the action of lime and carbon on ferrous sulphide at the high temperature of the arc furnace; (2) as calcium sulphide from the reaction of lime, ferrous sulphide, and calcium carbide at the higher temperatures of the arc furnace; (3) as calcium sulphide through the reaction of lime, ferrous sulphide, and silicon at the lower slag temperatures of the induction furnace; and (4) as iron sulphide from the action of ferrous oxide on calcium sulphide. The chemical reactions are as follows:

(1) FeS+CaO+C=Fe+CaS+CO.

(2) 3 FeS+2 CaO+CaC2=3 Fe+3 CaS+2 CO.

(3) 2 FeS+2 CaO+Si=2 Fe+2 CaS+SiO2.

(4) CaS+FeO=FeS+CaO.

The final reaction is reversible, acting from left to right in an atmosphere even slightly oxidizing, whereas the action is from right to left in the first three reactions.

From the equations it may be seen that in both the induction and the arc furnace the formation of a highly basic slag is essential, but desulphurization in the former, owing to its lower temperature, is brought about chiefly by the reducing action of silicon upon the oxides, whereas in the latter the reducing agent is carbon at a higher temperature.

"Ambery, R., Slags in electric steel refining: Metall, Chem. Eng., vol. 10, 1912, p. 601.

In the arc furnace complete deoxidation is recognized by the presence of calcium carbide in the slag. When deoxidation of the bath is complete the contents of the electric furnace represent the nearest approach to ideal equilibrium between different chemical compounds that has ever been accomplished in large metallurgical operations. In the converter and the open hearth the metals are under the action of air and gas, in the crucible the metal takes up carbon and silicon, whereas in the electric furnace the action of the metal on the basic lining is very slight and there is no exchange of elements between metal and slag, except that if the dephosphorizing slag of the oxidizing period has not been completely removed before deoxidization begins some of the phosphorus in this slag and some in the new slag will be reduced. This last point must be carefully watched in all refining in the electric furnace.

Although the induction furnace can not attain the high temperature necessary for desulphurizing by use of calcium carbide as can arc furnaces, the results obtained in refining practice are about equal.

COST OF PRODUCING STEEL IN THE ELECTRIC
FURNACE.

The cost of making steel in the electric furnace varies with local conditions. The cost of power does not enter so largely into the final cost as it does in some other electrometallurgical processes, especially the refining of molten steel. Plants are operating successfully under a power cost of 1 cent per kilowatt-hour in localities. where material can be obtained at the price common to other processes. Plants such as the one at Ugine, France, have been established in remote localities, where the cost of power is very low, 0.2 cent per kilowatt-hour, but the cost of material is high.

COST OF PRODUCING STEEL IN THE GIROD FURNACE.

Girod estimated in 1912 the cost of producing steel from cold scrap in the Girod furnace at Ugine as follows:

Cost of refining steel from cold scrap in Girod furnace at Ugine.

a

'Girod, P., The electric steel furnace in foundry practice: Metall. Chem. Eng., vol 10, 1912, p. 663.