Electric Arc Furnace Transformer: How to Size an EAF for a Steel Melt Shop, Why the Secondary Current in kA (Not the MVA) Decides the Design, and How Furnace Duty Differs From a Distribution Transformer

The answer first: an electric arc furnace transformer is sized from the melt shop's energy and power profile, and it is not a scaled-up distribution transformer. The kVA follows the melting power you need during meltdown, but three other numbers decide the design: the secondary current in kA (which sets the low-voltage winding, bushings, busbars and cooling), the tap range (which must be wide enough to keep the arc stable from cold scrap to a flat molten bath), and the impedance and series reactance (which set how hard the arc can short-circuit and how much flicker you export to the grid). Size it from kWh per tonne of steel and tap-to-tap time, specify the tap step table and the secondary current explicitly, and treat duty-cycle temperature rise, harmonic derating and repeated short-circuit withstand as the real acceptance criteria.

An electric arc furnace (EAF) transformer takes the voltage your mill receives — typically 33 kV, 66 kV, 110 kV or 132 kV, and 220 kV at very large integrated sites — and turns it into the few hundred volts that actually strike the arc: usually between about 400 V and 1,000 V, and up to 1,700 V in the largest modern designs. The currents on that side are the reason the machine looks nothing like a distribution transformer: 40 kA to 60 kA on a mid-size furnace, and more than 100 kA on the biggest units, with electrode currents up to about 120 kA in published furnace-transformer designs. What separates EAF duty from every other transformer application is that the load is a deliberately unstable moving short circuit, that the three phases are never balanced, and that the transformer is expected to be short-circuited thousands of times per year and still last 20 years.

Everything below is written for the person who has to buy or specify that transformer: a mill electrical engineer, an EPC contractor, or a scrap-based steel producer adding melting capacity. If you are sizing an ordinary distribution load instead, the general method in our sizing and budget formula guide and distribution system sizing guide is the right starting point; this article covers what changes when the load is an arc.

Key takeaways

• Energy first, kVA second. Melting energy (typically 350–500 kWh per tonne of liquid steel) divided by power-on time gives the average arc power; the average plus the meltdown peak gives the furnace-transformer rating. Modern ultra-high-power (UHP) furnaces are built around roughly 0.7–1.0 MVA of transformer per tonne of tap weight.
• The secondary current in kA is the binding constraint, not the MVA. The low-voltage winding, the copper-bar bushings, the water-cooled busbars and the electrode arms are all designed around a current that can exceed 100 kA.
• Furnace taps are not ±2 × 2.5%. An EAF transformer usually carries a wide tap range (a common arrangement is a secondary voltage band of roughly 2:1, for example 840 V down to 480 V) with 15–33 positions, often with primary taps at the neutral end and, for very wide ranges, an autotransformer or series-parallel arrangement.
• Harmonics and flicker are the grid-connection gate. The furnace must comply with the network operator's emission limits at the point of common coupling — IEC 61000-3-6 style harmonic limits and IEC 61000-3-7 flicker planning levels (Pst 1.0 / Plt 0.8 are the commonly quoted indicative values) — which usually means SVC or STATCOM compensation plus a filter, and a transformer specified for non-sinusoidal duty.
• Specify the duty, not just the nameplate. A furnace transformer order should state the tap table, the electrode short-circuit current, the duty-cycle temperature-rise test, the harmonic-derating basis, the number of short-circuit withstands expected and the melt-shop ambient temperature. Vague RFQs are how mills end up with a transformer that trips nuisance differential protection in year one.

What is an electric arc furnace transformer — and how is it different from a distribution transformer?

An EAF transformer is a special-duty power transformer whose secondary feeds the furnace electrodes through short, very heavy low-voltage connections. In an alternating-current (AC) furnace there are three graphite electrodes that descend into the scrap; the arc strikes from each electrode to the charge, and because the delta or star point of the arc is the molten bath itself, the electrical circuit changes second by second as the scrap melts, the electrodes burn away and the operator changes tap. Published furnace-transformer ranges run from about 10 MVA to 300 MVA, with secondary voltages up to 1,700 V and electrode currents up to about 120 kA.

The differences from a distribution transformer are not cosmetic. They are structural:

Two consequences follow directly from that table. First, you cannot buy an EAF transformer from a distribution-transformer catalogue: it is engineered per project, and the engineering is yours to specify. Second, the acceptance tests that matter are the special ones — not the routine ratio and impedance checks, but the duty-cycle temperature-rise test, the demonstration of short-circuit withstand, the measurement of load loss at each tap and the sound-level measurement on a unit that will sit next to the melt shop.

What does a steel melt shop actually demand from its transformer?

A scrap-based melt shop runs in heats. Each heat is a repeating cycle: charge scrap (often two or three buckets), bore-in, meltdown, refining and alloying under a slag, then tapping into a ladle; in many mills a separate ladle furnace then reheats and refines the steel before casting. For a typical 100-tonne AC furnace the cycle is about 40–60 minutes, of which perhaps 30–40 minutes is power-on time, and about 30–35 heats are produced per day.

The electrical profile inside one heat has a shape, and the transformer has to follow it:

• Bore-in (first minutes). The electrodes descend onto cold, uneven scrap; the arc is unstable and short circuits constantly. Currents swing violently and voltage flicker is worst here.
• Meltdown. The furnace runs at, or close to, maximum available arc power for most of the power-on time. This phase sets the transformer rating.
• Refining. The bath is flat and the arc is stable; power is reduced to roughly 50–65% of maximum to control the bath temperature and slag.
• Idle. Charging, tapping and delays — no load on the furnace transformer, but the transformer is commonly de-energized and re-energized between melts.

Three design facts fall out of that duty:

1. The transformer is thermally cycled, not thermally steady. Loading swings between zero and full melt power several times an hour, all day. That cyclic duty, not continuous nameplate loading, is what the transformer's oil and winding temperatures follow, and it is why a duty-cycle temperature-rise assessment matters (see load factor and overload capacity for how cyclic loading is evaluated).
2. The transformer must be re-energized frequently. Every heat begins with a switching operation, which means every heat begins with an inrush current event. Differential protection must be inrush-restrained or it will misoperate.
3. Ambient is hotter than a substation hall. A furnace transformer normally shares a building or an adjacent bay with the furnace itself, where air temperature can sit well above 40 °C. Derating for ambient (see derating practice) belongs in the calculation, not in a footnote.

Step 1 — How much electrical energy does one tonne of steel need?

Start from energy, not from a kVA-per-tonne guess. The electrical energy needed to melt a tonne of steel scrap depends mainly on scrap quality, charge mix (scrap only, or scrap plus direct-reduced iron and pig iron), oxygen and carbon injection practice, and whether the off-gas is used to preheat the scrap.

That single number — kWh per tonne — is the anchor for everything else. It is also the number that a financing model cares about, because at an industrial electricity price of, say, USD 0.07/kWh, every 50 kWh/t of extra consumption is roughly USD 3.5 per tonne of steel, which on a 500,000 t/year mill is more than USD 1.7 million a year. Getting the transformer and the arc-control philosophy right is therefore an energy-cost decision as much as a hardware decision.

Step 2 — How does melting energy become a transformer MVA?

Work the arithmetic in four moves. The illustration below is deliberately round-number and uses a 100-tonne furnace; substitute your own mill data.

1. Energy per heat. 100 t × 400 kWh/t = 40,000 kWh (40 MWh) of electrical energy per heat.
2. Power-on time. With a 45-minute tap-to-tap cycle and 60–70% power-on time, the arc is on for about 0.5–0.53 hours.
3. Average arc power. 40,000 kWh ÷ 0.53 h ≈ 75 MW average during power-on. If your practice is more efficient (350 kWh/t with preheating) or your heat is longer, the average drops; an 8-minute-longer power-on time with the same energy gives about 60 MW.
4. Meltdown peak and transformer rating. The transformer must deliver the meltdown (maximum) arc power, which is higher than the cycle average. Arc-circuit power factor is typically in the range of about 0.75–0.85, so the furnace transformer's apparent power is the peak active power divided by that power factor. A furnace that must sustain roughly 70 MW of arc power at a power factor around 0.8 needs about 85–90 MVA at the furnace-transformer terminals.

The last column is the one to sanity-check against industry practice: modern ultra-high-power furnaces are specified at roughly 0.7–1.0 MVA of furnace-transformer rating per tonne of tap weight, which bands a 100-tonne furnace between about 70 and 100 MVA. If your arithmetic puts you materially outside that band, one of your inputs — energy per tonne, power-on time or power factor — is off, and the project is worth re-checking before it reaches an RFQ.

Step 3 — Why does the secondary voltage decide everything else?

Once you know the apparent power, the secondary voltage fixes the current, and the current fixes the hardware. The relation is the ordinary three-phase one:

I₂ = S ÷ (√3 × U₂)

Put 90 MVA through an 840 V secondary and you get about 62 kA per phase. Drop the voltage to 480 V and the same current delivers only about 50 MVA — which is exactly the behaviour a furnace operator wants, because the electrode current is the constraint, and the design keeps it roughly constant while the voltage changes. An arc furnace at a low voltage with the same current is running a short, hot arc with a lower arc power; at a high voltage it runs a long arc with a higher arc power but more radiation and more refractory wear. Choosing the operating point is metallurgy; delivering it is the transformer's tap design.

What the kA number drives in the design

• The low-voltage winding. At 50–120 kA, the secondary is not a wire winding. It is usually a small number of turns of heavy copper bar or sheet, arranged to keep eddy losses and stray losses manageable, with carefully calculated clamping because electromagnetic forces during a short circuit rise with the square of the current.
• The bushings. Low-voltage connections are made with copper-bar bushings — often several in parallel — and published furnace designs switch from bringing the low-voltage out through the tank cover (below about 80 kA) to bringing it out through the tank wall at the highest currents, so that the bars reach the electrode arms on the shortest possible path. For the general principles see our bushing specification guide.
• The secondary busbars and electrode arms. These are water-cooled tubes or welded bar assemblies sized for current density and for mechanical stiffness, not for a voltage drop that anyone worries about — at 60 kA a single milliohm of connection resistance is 3.6 kW of loss in the circuit and a genuine thermal problem.
• The tap table. Because the current stays high while the voltage falls, the tap changer must break and make at the full secondary current without allowing the arc to go out. This is why furnace tap changers are on-load, and why their step count and mechanical endurance are specified separately from those of a network transformer (see OLTC vs off-circuit tap changers).

Step 4 — Why is the tap range so much wider than a distribution transformer's?

A distribution transformer's off-circuit tap changer exists only to trim for system voltage: five positions, typically ±2 × 2.5%. A furnace transformer's tap changer is a power-control device. It sets the arc voltage and therefore the arc length and the power delivered to the bath, heat after heat, from cold scrap to a flat molten bath.

Three mechanisms make that wide range possible, and the right choice for your mill depends on how wide the range must be:

1. Taps at the neutral end of the primary winding (“variable flux” regulation). This is the most compact and economical arrangement for the usual regulation ranges. Because the number of primary turns in circuit changes, the core flux changes with the tap, so the secondary voltage — and the kVA the transformer can deliver — falls together as you go down the tap range, with the secondary current staying close to constant.
2. Regulation with an autotransformer. When the required voltage range is wider than a straightforward tapping winding can cover economically, a regulating autotransformer is inserted in series with the furnace transformer's primary. This is the classic way to reach a very wide range while keeping the furnace transformer itself simpler.
3. Series-parallel, delta-star or reversing tap arrangements. Special tap-changer connections change the winding configuration rather than just the turns in circuit, which extends the ratio range further; the price is a more complex, more expensive changer with fewer positions.

Practical consequences for the specification:

• Ask for the tap table, not a tap percentage. The deliverable is a table of secondary voltages per tap position, the corresponding electrode current and the guaranteed impedances. Furnace transformers are commonly specified with something like 15–33 positions, and the load-loss guarantee must be given at each tap, not only at the rated tap.
• Expect on-load operation. In a modern furnace the tap changer moves during melting as the regulator hunts for the best arc stability; that is a mechanical endurance requirement measured in hundreds of thousands of operations, so it should be stated and witnessed. Our voltage-regulation guide explains the underlying mechanics.
• Decide whether you need per-phase regulation. Because the three arcs never share load equally, published furnace-transformer practice provides separate voltage regulation for each phase (and therefore separate single-phase switching) when the asymmetry is significant. If your operating philosophy relies on setting one phase differently from the others, say so in the specification — it changes both the transformer and the circuit breaker arrangement.

Step 5 — What is impedance for in an arc furnace, and when do you need a series reactor?

In a network transformer, impedance is something to be minimised so that voltage regulation is good. In a furnace circuit, impedance is a control parameter. The arc is an unstable, negative-resistance load; a certain amount of circuit reactance is what keeps it stable, limits how violently the current rises when an electrode touches the scrap, and converts the arc into a controllable power source. Two distinct things provide that reactance:

• Transformer impedance. Furnace transformers are usually specified with a higher short-circuit impedance than a network transformer of the same rating — commonly in the region of 10–14% at base rating, with high-reactance designs going higher. That extra reactance is bought deliberately: it costs efficiency and voltage regulation, and it buys arc stability and lower flicker.
• A series reactor. In larger installations the required reactance is provided or topped up by a reactor connected in series with the furnace transformer — either in a separate tank or integrated in the same tank as the transformer. These reactors are themselves tapped, and their reactance can be varied from essentially zero up to their full value using no-load or on-load tap changers, which lets the mill tune the arc-circuit characteristics at commissioning and re-tune them when the scrap mix or product mix changes.

Design consequences: an EAF transformer's low-voltage short-circuit current is not a fault to be minimised but a rated design condition, and the electrode short-circuit current is a nameplate value to be specified. Typifying a number is the buyer's job, not the manufacturer's; a mill that states “short-circuit at the electrodes shall not exceed X kA at the lowest tap” gets a machine that matches its electrode and control system. Our short-circuit current calculation guide shows how the reactance budget combines transformer impedance, series reactor, busbars and cables.

Step 6 — Which vector group does a furnace transformer use, and why is the secondary delta?

Furnace-transformer connections are chosen for three reasons: to reduce the current each winding has to carry, to block triplen harmonics from circulating, and to leave the arc circuit free of a neutral point. The common answers:

• Delta-connected secondary. In a delta the winding current is 1/√3 of the line current, so a 60 kA line current becomes about 35 kA per winding — a large reduction in copper, in losses and in clamping forces. Delta also gives triplen (3rd, 9th, 15th) harmonic currents a closed path so they do not leave the transformer, and it means there is no neutral terminal — the circuit's star point is the molten bath itself.
• Primary directly connected to the incoming supply and tapped at the neutral end for the usual regulation range, or supplied through a regulating autotransformer for very wide ranges.
• Vector group. Furnace transformers are usually named as Dd0-type (delta–delta) units. Where a slightly different arrangement is required for protection or for connection to an existing installation — for example a star primary with an available neutral — a Ynd-type vector group may be used instead; what matters is that the group is stated explicitly and that the protection scheme is designed around it. Our vector group guide covers how the symbol maps to connection and phase shift, and our parallel operation guide covers the conditions that must be met if a spare furnace transformer may ever have to run in parallel with another unit.

Because each phase of a furnace load is unequal, the transformer must be designed for uneven phase currents and for the resulting negative-sequence and zero-sequence components in the supply. Say so in the specification: it affects winding design, tap-changer rating and the choice of the furnace circuit breaker.

How much harmonic duty does an EAF put on its transformer?

During bore-in and meltdown, the arc is a highly non-linear load and the current waveform is heavily distorted — the second and third harmonics usually dominate, with higher-order and interharmonic components present as well. Once the bath is flat the distortion falls substantially, but the melting phases are what the design must survive.

Two separate obligations follow:

1. Compliance at the point of common coupling. The network operator will apply harmonic emission limits (in the IEC framework, the assessment of distorting loads in MV and HV systems, aligned with the compatibility levels the installation is expected to respect) and, in many countries, IEEE Std 519-style voltage-distortion limits are written into the grid code. In practice a large EAF cannot meet these limits on its own: the melt shop is supplied with harmonic filters and a reactive-power compensator — typically a TCR/FC static var compensator, sometimes a STATCOM — and the transformer, the SVC transformer and the filter duties are designed together. Treat the SVC and the furnace transformer as one electrical package; specifying them separately is how mills end up with a compensator that cannot control the distortion the furnace produces.
2. Capability of the transformer itself under non-sinusoidal current. Harmonic currents raise winding eddy losses and stray losses well beyond what a sinusoidal test tells you, which is precisely the situation the derating practice for supplying non-sinusoidal loads exists to cover. Practically, this means specifying the harmonic spectrum and duty the transformer must carry, and either deriving a K-factor / derating requirement from it or stating the temperature-rise basis under distorted loading. Our harmonics guide and K-factor selection guide explain the mechanism and the numbers; the important point is that an EAF transformer is not assessed as if it carried a clean sinusoidal current.

Why is flicker the limit that decides whether a new furnace can connect?

Voltage flicker — the rapid, repeated voltage fluctuation that a furnace produces every time an electrode touches scrap — is usually the constraint that determines how much furnace power a network operator will allow at a given connection point, and whether compensation is required. It is measured as the short-term severity index Pst and the long-term index Plt, and the IEC framework addresses the connection of fluctuating installations to MV, HV and EHV systems with indicative planning levels that are commonly quoted as Pst = 1.0 and Plt = 0.8. Utilities often publish their own, tighter values.

What a mill should know before it specifies hardware:

• Flicker depends on the short-circuit capacity of the network you connect to. The same furnace is far more disturbing on a weak rural 33 kV system than on a stiff 220 kV busbar. This is why the same furnace design can be acceptable at one site and refused at another, and why the flicker study has to be done for your connection point, not for the furnace alone.
• Furnace design changes flicker. High-reactance furnace designs — more circuit reactance, whether in the transformer or in a series reactor — reduce flicker, and published flicker-assessment guidance uses correction factors of roughly 0.8–1.0 for high-reactance furnaces relative to a standard design. Direct-current (DC) arc furnaces typically show lower flicker severity for the same tonnage, which is one of the reasons the DC route is chosen on weak grids.
• Compensation is usually mandatory anyway. Most new melt shops are required to install an SVC or STATCOM sized to keep flicker within limits at the point of common coupling, and the transformer order should state whether the compensator is on the same busbar, downstream, and what the transformer must do when the compensator trips.

Ask for both halves of the number when you negotiate the connection: how much furnace power is permitted, and on what flicker basis. A mill that buys a transformer for full melt power and then discovers the connection is limited to 70% of it has bought a machine it cannot use.

Why must a furnace transformer survive repeated short circuits and repeated energization?

An arc furnace transformer endures more fault duty than almost any other transformer of its size, because faults are part of normal operation:

• Repeated short circuits. Every time an electrode contacts the scrap, the secondary is short-circuited through the arc, and the furnace circuit breaker may or may not operate — in many designs the regulator simply backs the electrode off. The transformer therefore sees thousands of short-circuit events per year. The applicable requirement is the standard short-circuit withstand capability that power transformers must be shown to satisfy, evaluated over the duration and with the peak current the design practice assumes; for a furnace transformer it is worth stating the number of withstands you expect and asking for the mechanical design (clamping, spacers, winding support) to be justified against it, not merely analysed once.
• Repeated energization. Once per heat, the furnace breaker closes and the transformer is energised again — sometimes with the core still retaining residual flux from the previous melt. Inrush current can reach several times rated current with a heavily asymmetric waveform, so the differential protection must be inrush-restrained and the transformer's mechanical design must tolerate the repeated electromagnetic shock. See our inrush current guide.
• Winding movement accumulates. Each short circuit produces a mechanical force proportional to the current squared, and it acts thousands of times. Manufacturers who understand furnace duty design for it explicitly; the RFQ should ask the question.

How does cooling and ambient differ from a normal substation?

Furnace transformers almost always need more cooling per MVA than a network transformer, and they need it in a hotter, dirtier place. Expect:

• Forced cooling as the normal mode. ONAN (oil natural, air natural) alone is rarely enough at furnace ratings; the typical arrangement is oil-forced, air-forced (OFAF), with fans and oil pumps running during meltdown.
• Water-cooled low-voltage components. The secondary busbars, and in some designs the windings themselves, are water-cooled; the cooling-water supply and its redundancy become part of the plant's critical infrastructure, not an auxiliary detail.
• A higher assumed ambient. A melt-shop electrical bay commonly runs above 40 °C because the building is dominated by the furnace. Insisting on a realistic ambient temperature in the guarantee avoids a transformer whose nameplate cannot be achieved in its actual location.
• Duty-cycle temperature rise. Because loading is cyclic, the meaningful thermal assessment is a duty-cycle one: a documented load cycle, a defined ambient, and a demonstrated hot-spot temperature that keeps insulation ageing acceptable over the melt cycle (see cooling methods and temperature rise and insulation class). The temperature-rise test for a furnace transformer is normally specified with the furnace load cycle, not at continuous rated load only.
• Noise. Melt shops are noisy places, but the transformer's tank, coolers and the series reactor still have to meet the mill's environmental permit, and a harmonic-rich load will make the core noisier than a clean load. Set the sound-pressure level and the measurement standard in the specification (see noise level by kVA).

What protection and monitoring does a furnace transformer need?

Furnace protection has to be fast and stable at the same time: fast, because a short circuit at the electrodes damages the transformer and the electrode arms quickly; stable, because nuisance tripping stops production. The conventional package is:

• Transformer differential protection with harmonic (inrush) restraint, plus overcurrent and earth-fault backup on the furnace breaker.
• Sudden-pressure / pressure-relief and Buchholz-type gas protection to catch an internal fault that raises gas or oil velocity before the differential element sees it. Our protection devices guide explains what each device detects and how to specify it.
• Oil and winding temperature indication and tripping, with the alarm levels set for furnace ambient, plus a cooling-control philosophy that starts the pumps and fans before the heat begins.
• Dissolved-gas analysis on an accelerated schedule. A furnace transformer accumulates fault energy far faster than a distribution unit; quarterly sampling during the first year, and continuous gas monitoring on critical units, is normal. Our DGA guide covers interpretation.
• Series-reactor protection where a reactor is fitted, and coordination with the SVC/filter protection so that a compensator trip does not turn into a furnace trip.

Before energisation, the commissioning programme should include the standard pre-energisation checks adapted to furnace duty (see commissioning tests checklist), plus verification of the tap table by ratio measurement at every position, and a functional test of the arc regulator operating the tap changer and electrodes together.

What else in a scrap-based mill needs a transformer?

The furnace transformer is the headline, but a melt shop is a transformer-rich plant, and buying them as one coordinated package usually beats buying them one at a time:

• Ladle furnace transformer. Same technology family, lower secondary voltage and very high current, used under a slag for refining; its duty is more continuous but still unbalanced.
• Rectifier transformer for a DC furnace (or for a large DC drive elsewhere in the mill), where the transformer feeds a thyristor rectifier and must be designed as a converter transformer, with the pulse number of the rectifier driving the harmonic spectrum and the rating.
• Melt-shop distribution transformers for cranes, conveyors, dust extraction, water treatment, oxygen and compressed-air plant, and the electrode-hydraulic and control systems.
• SVC or STATCOM transformer connecting the compensator and filters to the busbar, plus filter reactors and capacitors.
• Continuously cast and rolling-mill drives downstream, whose own harmonic and load-fluctuation duty should be part of the same power-quality study, so the compensation is sized once for the whole plant rather than twice for its parts. Where motors are a dominant load, the general method in our distribution system sizing guide and sizing formula guide applies, with voltage class selection deciding the busbar voltages.

What drives the price and the lead time of an EAF transformer?

Furnace transformers are quoted per project, so treat any catalogue price with suspicion. The cost drivers, in rough order of influence:

1. MVA and secondary current. The kA at the secondary side is the single most expensive parameter, because it sets copper volume, bushing count, water-cooling and clamping.
2. Tap-changer complexity. Step count, current rating, reversing or series-parallel arrangements and per-phase regulation all add cost and delivery risk, because the tap changer is a bought-in component with a long lead time.
3. Cooling scope. Oil pumps, fans, water-cooled busbars and their controls; whether the series reactor shares the transformer tank.
4. Winding material. Copper, aluminium or copper-in-tank-alu-in-LV hybrids; the choice interacts with losses, temperature rise and capital cost (see total cost of ownership and oil preservation systems).
5. Transport and site. Above a certain size the transformer cannot move as one piece on normal routes, so it is built as three single-phase units or designed to be split, which changes both price and the site work. Site access, allowable transport weight and the need for a spare unit all belong in the RFQ (see transportation and installation and warranty terms).

Lead time follows the same order. Availability of the tap changer and the time needed for a duty-cycle temperature-rise test and a short-circuit withstand demonstration are usually the critical path, not the winding work. Published lead-time bands for large power and special-duty transformers of this class run into many months, and the practical planning rule for a furnace transformer is to start the RFQ at the same time as the furnace order, not after it (see delivery lead time).

How to specify an EAF transformer: the RFQ parameter table

Hand the manufacturer the following, and you remove most of the ways an EAF transformer order goes wrong:

Six specification mistakes that cost mills money

1. Buying on MVA alone. The secondary current and the tap table are what the furnace actually needs; a machine that meets the MVA but not the current cannot melt your scrap at the power you planned.
2. Assuming a standard vector group and standard impedance. Furnace duty wants a delta secondary, a higher impedance and usually a series reactor. Defaulting to a network-transformer design produces an unstable arc and heavier flicker.
3. Leaving harmonic derating and flicker to the supplier. Both are system issues. The mill owns the harmonic spectrum and the connection-point limits; the manufacturer can only design to what you state.
4. Specifying a continuous temperature-rise test for a cyclic load. A duty-cycle test with your melt profile and your ambient is the number that predicts service life.
5. Forgetting the spare. A single furnace transformer with no spare is a single point of failure for the entire plant; the cost of a spare should be compared against the value of lost production, not against the transformer price (see TCO and supplier audit checklist).
6. Skipping factory acceptance. For special-duty transformers, witnessing the ratio test at every tap, the loss measurement and the temperature-rise test at the factory is far cheaper than discovering a tap-table error during commissioning (see pre-shipment inspection).

How QDTB helps you specify and price it

QDTB is an engineering and supply partner for power and distribution equipment, with our own plant base in Qingdao, Shandong. For melt-shop projects we work from the mill's process data upward: energy per tonne and heat cycle first, then furnace-transformer rating, tap table, impedance and reactor requirement, then the surrounding package — incoming HV/MV switchgear, melt-shop distribution transformers, the compensator and filter transformer, and the LV busbar and electrode-side hardware. If you want a starting point before an engineer is assigned, our engineering toolbox calculator produces equipment lists and budget-level numbers from the same load inputs, and the power system configurator builds a system architecture you can hand to a designer. For a furnace-specific enquiry, send the furnace size, tap-to-tap time, energy per tonne, incoming voltage and the connection-point short-circuit level, and we will come back with a sizing basis and an indicative FOB Qingdao price for the transformer and the reactor.

Sources / 资料来源

• Source: IEC 60076 series — Power transformers: Part 1 (General), Part 2 (Temperature rise), Part 3 (Insulation levels and dielectric tests), Part 5 (Ability to withstand short circuit), Part 7 (Loading guide for oil-immersed transformers), Part 10 (Determination of sound levels), Part 11 (Dry-type transformers), Part 20 (Energy efficiency)
• Source: IEC 61378-1 — Converter transformers for industrial applications (applies to rectifier transformers feeding DC arc furnaces and large drives); IEC 61378-3 — Application guide
• Source: IEEE Std C57.18.10 — Standard practices and requirements for semiconductor power rectifier transformers
• Source: IEC 60214-1 — Tap-changers, Part 1: Performance requirements and test methods (on-load and off-circuit tap changer duty and endurance)
• Source: IEC 60076-18 — Measurement of frequency response (FRA)
• Source: IEC 60076-22 series — Power transformer and reactor fittings, including protective devices
• Source: IEC 60909 series — Short-circuit currents in three-phase a.c. systems
• Source: IEC 61000-3-6 — Assessment of emission limits for distorting loads in MV and HV power systems; IEC 61000-2-4 — Compatibility levels for low-frequency conducted disturbances
• Source: IEC 61000-3-7 — Assessment of emission limits for the connection of fluctuating load installations to MV, HV and EHV power systems; indicative planning levels for flicker commonly quoted as Pst = 1.0 and Plt = 0.8
• Source: IEC 61000-4-15 — Flickermeter: functional and design specifications
• Source: IEEE Std 519 — Recommended practice and requirements for harmonic control in electric power systems (voltage distortion limits at the point of common coupling)
• Source: IEEE Std C57.110 — Recommended practice for establishing liquid-immersed and dry-type power transformer capability when supplying nonsinusoidal load currents (harmonic derating)
• Source: IEEE Std C57.91 — Guide for loading mineral-oil-immersed transformers and step-voltage regulators
• Source: EPRI — Guidelines for flicker limits at industrial customers: flicker planning levels in MV, HV and EHV systems; the furnace-type correction factor of about 0.8–1.0 for high-reactance furnace designs; lower flicker severity constant (Kst) typical of DC arc furnaces
• Source: published arc-furnace flicker studies applying IEC 61000-3-7 indicative planning levels (Pst 1.0 / Plt 0.8) to steel-plant installations
• Source: SEL (Schweitzer Engineering Laboratories) technical paper — Protecting electric arc furnace transformers (protection practice, and series-reactor adjustment as a means of modifying the available short-circuit current)
• Source: GE Vernova / GE Grid Solutions — Arc furnace power transformers brochure: published furnace-transformer ratings of 10–300 MVA, secondary voltages up to 1,700 V, electrode currents up to 120 kA, delta-connected secondary preferred to reduce the current carried, taps at the neutral end of the primary winding (variable-flux regulation) as the most compact and economical arrangement, regulation with an autotransformer for wide ranges, in-tank or separate series reactors with reactance adjustable from 0 % to 100 % by no-load or on-load tap changers, copper-bar bushings with low-voltage output through the tank cover below about 80 kA and through the tank wall at the highest currents, and separate voltage regulation per phase because of asymmetric three-phase loading
• Source: World Steel Association (worldsteel) — World Steel in Figures 2026 and monthly crude steel production statistics: world crude steel production of about 1.85 billion tonnes in 2025, including 960.8 Mt in China; Short Range Outlook October 2026: global steel demand essentially unchanged at 1,722 Mt in 2026 before returning to growth in 2027
• Source: OECD Steel Outlook 2025 — electric arc furnaces produced 28% of world crude steel in 2023 (citing World Steel Association data); blast furnace/basic oxygen furnace production accounts for about 70% of world production; electric arc furnace projects account for nearly 60% of steelmaking capacity under way or planned over the three years from 2025
• Source: published EAF energy-consumption studies — electrical energy of roughly 350–500 kWh per tonne of liquid steel in conventional practice, about 280–400 kWh/t at best-in-class plants with scrap preheating and oxygen injection, against a theoretical melting minimum of roughly 300 kWh/t
• Source: GB/T 6451 — Technical parameters and requirements for three-phase oil-immersed power transformers; GB 20052 — Minimum allowable values of energy efficiency and energy efficiency grades for power transformers
• Source: ISO 12944 series — Corrosion protection of steel structures by protective paint systems, including corrosivity categories applicable to hot, dusty industrial buildings; ISO 1461 — Hot dip galvanized coatings on fabricated iron and steel articles
• Source: QDTB Engineering Team — industry experience / 行业经验值 for the 0.7–1.0 MVA furnace-transformer rating per tonne of tap weight band, arc-circuit power factor range of about 0.75–0.85, furnace-transformer short-circuit impedance commonly around 10–14% at base rating, melt-shop ambient temperatures above 40 °C, per-phase regulation practice, water-cooled busbar and electrode-arm practice, and the cost and lead-time ranking of MVA, secondary current, tap-changer complexity, cooling scope and transport constraints