Electric Arc Furnace

Arc performance depends on two coupled variables: arc voltage and current. For each arc length there is only one possible voltage, and the power delivered results from the combination of both across the three phases.

Arc voltage is proportional to the distance between the electrode tip and the metallic charge. The regulator does not control that distance directly: it controls current. If a movement of the charge shortens the arc, current rises, the voltage drop across the reactance increases, arc voltage falls and the regulator raises the electrode until the original point is restored. It is current-based control logic, not position-based.

In a three-phase furnace, arc voltage is determined by the supply voltage and by the phase shift between voltage and current, and that phase shift is in turn set by the current and the circuit reactance. In other words: the operating point is not freely chosen; it is constrained by the real impedance of the installation.

The arc is not a linear load: the current waveform is not sinusoidal. Every time the current crosses zero —120 times per second in a 60 Hz system— the arc extinguishes and must reignite. That reignition delay generates harmonic currents that circulate through the circuit and raise the effective reactance.

The practical consequence is that the real reactance of the furnace circuit is noticeably higher than that calculated with a sinusoidal waveform. The ratio between the two is known as the operating factor; it is higher during scrap melting than during refining with foaming slag, and it grows with the sinusoidal cos φ.

The effect on the result is large: the real arc voltage and active power end up well below what the sinusoidal calculation would give. In practice, the real power factor rarely exceeds values on the order of 0.70 at the start of melting, 0.76–0.77 at mid-melt and 0.82–0.85 during refining.

There is a threshold below which the regulator loses control of the arc. The condition for stability is that a decrease in current produces an increase in arc voltage. Above a certain sinusoidal cos φ that relationship reverses: as current drops, arc voltage stops rising, stalls and then falls sharply.

When you try to operate below that limit, the regulator enters a pumping cycle: it raises the electrode trying to lower the current, the arc extinguishes, it orders a lowering, the current spikes, it raises again. The result is a violent drop in power, an increase in specific consumption and severe vibration of arms and electrodes that, in extreme cases, causes breakage.

The threshold is not the same throughout the heat: the most turbulent stage of melting is the most exposed, while during refining with foaming slag the stability margin is considerably larger. Where that threshold falls in a specific installation depends on its real impedance and can only be determined by calculating it.

Stability and arc length are independent properties. An arc can be more than a meter long and perfectly stable, as long as voltage, current and reactance place operation within the stable zone. In fact, at equal power factor, a long arc tends to be more stable than a short one, because the length variations caused by charge movement are proportionally smaller.

Nor is it correct to associate a long arc with low current. Lowering current lengthens the arc only slightly, and if it is lowered below the stability limit the effect reverses: the arc becomes shorter, not longer. It is a common and costly confusion, because whoever makes it believes they are opening the arc while they are actually extinguishing it.

Refractory erosion is assessed using a heat radiation index that combines arc power and arc voltage, divided by the square of the distance between the electrode and the wall. The quadratic dependence on distance explains why a long, uncovered arc punishes the lining so much.

For the same arc power there are several possible operating points, and they are not equivalent: some favor electrode life and others favor lining life. The choice is a trade-off with measurable economic consequences in both directions, and it depends on the real circuit impedance, the furnace design and what is constraining the plant. It is a decision made with your own numbers, not with a general rule.

Foaming slag resolves much of the conflict: by burying the arc it improves energy transfer —on the order of 20 kWh/t— and at the same time protects walls and roof. It is the same phenomenon read in two different metrics: kWh/t and refractory wear.

Electrode consumption has two components: tip consumption, which is related to the ratio between current and arc voltage, and side oxidation consumption, which depends on the exposed surface and the square of the current.

This leads to a counterintuitive warning: reducing current does not always reduce consumption. When current is too low relative to the electrode diameter, a further reduction increases it, and the effect worsens below the stability limit. That is why electrode consumption is tackled from the electrical operating point, not from the purchase order.

Typical diameters in 50 to 150 t furnaces are 500, 550 and 600 mm, with 700 mm and occasionally 800 mm also manufactured. Reference consumption is between 1.0 and 2.0 kg/t, trending lower in UHP grades.

Foaming occurs when injected carbon reacts with the oxygen dissolved in the bath and with the FeO in the slag, generating carbon monoxide. The CO bubbles pass through the slag and expand it.

For foaming to be effective, slag viscosity must exceed a certain threshold; below that value the bubbles pass through without expanding it. Slag chemistry —basicity, FeO and MgO content— is therefore an energy control variable, not just a metallurgical one.

Refining in the furnace includes deoxidation and desulfurization, plus control of nitrogen, hydrogen and oxygen. Deep decarburization and inclusion metallurgy belong to secondary refining.