Melting Furnace

A melting furnace is the heart of any foundry or steel plant: the unit that converts a cold charge of metal scrap, ingot, or returns into homogeneous liquid metal at a controlled temperature and chemistry. The category spans an enormous range, from a 5 kg bench crucible furnace melting silver, to a 420 tonne electric arc furnace recycling steel scrap, with melting points ranging from 660 degrees Celsius for aluminum to around 1,500 degrees Celsius for steel, tapped with superheat above 1,600 degrees Celsius.

Because no single design covers that range, engineering selection is the act of mapping metal type, batch tonnage, chemistry tolerance, and energy source onto a specific furnace technology: induction, electric arc, cupola, crucible, or reverberatory. This guide decodes those technologies, their key specifications, and the standards a procurement engineer should verify before a six- or seven-figure capital decision.

White-hot molten steel pours from a 35-ton electric melting furnace at the Allegheny Ludlum steel plant, showing the furnace shell, electrode platform, and glowing liquid metal stream

This guide is written for foundry, metallurgical, and procurement engineers. Across six chapters it covers furnace classification, induction and arc working principles, refractory linings and crucibles, capacity and power specifications, spec-sheet decoding, and selection decisions, with seven selection FAQs and maker comparisons. Parameters reference the IEC 60519 series on safety in electroheat installations, NFPA 86 on ovens and furnaces, and published manufacturer data from Inductotherm, ABP Induction, Otto Junker, and major steel-plant builders.

Chapter 1 / 06

What is a Melting Furnace

A melting furnace is an industrial thermal device that raises a metallic charge above its liquidus temperature, holds it as a liquid bath, and delivers it at a tapping or pouring temperature suitable for casting. It differs from a heat-treatment or annealing furnace, which keeps metal solid below its melting point, and from a forging furnace, which heats billets to a plastic but solid state. The defining duty of a melting furnace is a controlled solid-to-liquid phase change followed by superheat: the bath is taken a defined number of degrees above the melting point so the metal stays fluid through transfer, pouring, and mold filling.

Functionally, every melting furnace combines four subsystems: an energy source that injects heat (electrical induction, electric arc, combustion of coke or gas, or electrical resistance); a refractory containment that holds the liquid without dissolving or leaking; a charging and tapping arrangement that gets cold metal in and liquid metal out; and an instrumentation layer that measures temperature and, on advanced units, infers chemistry. The performance of these four together determines melt rate, energy use per tonne, metal loss, and chemistry consistency.

The history of controlled metal melting runs from the ancient clay-graphite crucible, through the coke-fired cupola that industrialized cast-iron production in the eighteenth and nineteenth centuries, to the electric era. The electric arc furnace was commercialized by Paul Heroult in 1900 and became the dominant route for recycling steel scrap. The coreless induction furnace, whose modern medium-frequency form was matured through the twentieth century, became the workhorse of the foundry floor. ABP Induction traces its first production induction furnace to 1903, and the Inductotherm Group, founded in 1953, has since supplied over 36,500 systems to more than 130 countries, an indication of how thoroughly induction melting displaced older methods in clean-melting applications.

In scale terms the category spans roughly five orders of magnitude of batch size. A jeweller's tilting crucible furnace melts a few hundred grams to a few kilograms of gold or silver. A typical iron foundry runs coreless induction furnaces of 0.5 to 12 tonnes. A bulk steel mill operates an electric arc furnace of 80 to 200 tonnes, and the largest scrap-only DC furnace on record, operated by Tokyo Steel in Japan, taps about 420 tonnes per heat, fed by eight 32 MVA transformers for roughly 256 MVA of installed power. The same word, furnace, covers all of these, but almost no engineering detail carries across them unchanged.

Four engineering metrics anchor every comparison in this guide: melt rate (tonnes or kilograms per hour), specific energy consumption (kWh per tonne, or kg of coke per tonne for a cupola), refractory campaign life (heats or weeks before relining), and chemistry control (carbon, sulphur, and alloy pickup or loss). These four, weighted by local energy price and labour, set the lifetime cost of a melting installation far more than the headline purchase price does.

Chapter 2 / 06

Furnace Types and Classification

Melting furnaces are classified first by energy source (electric versus fuel-fired) and then by how that energy reaches the metal. The five mainstream industrial families are coreless induction, channel induction, electric arc, cupola, and fuel-fired crucible or reverberatory. Each occupies a distinct niche defined by tonnage, metal type, and chemistry tolerance, and choosing the wrong family is the costliest mistake in a melt-shop design. The table below summarizes their operating envelopes.

Furnace TypeTypical CapacityMetalsEnergy SourceBest For
Coreless induction0.005 to 30 tIron, steel, Cu, AlMedium-freq inductionBatch melting, alloy control
Channel induction8 to 120 tCu, brass, Al, ironLine-freq inductionHolding, superheating
Electric arc (EAF)10 to 420 tCarbon and alloy steelGraphite-electrode arcBulk scrap recycling
Cupola1 to 100 t/hGrey and ductile ironCoke combustionContinuous iron melting
Crucible / reverberatory0.05 to 50 tAl, Cu, Zn, preciousGas, oil, or resistanceNon-ferrous batch melting

Coreless induction furnaces are the foundry workhorse. A water-cooled copper coil surrounds a refractory crucible, and an alternating magnetic field induces eddy currents directly in the charge, heating and electromagnetically stirring it. They melt a cold solid charge from empty, switch alloys quickly, and give low melt loss and clean operation, which is why they dominate iron, steel, and non-ferrous batch melting from a few kilograms to about 30 tonnes. The trade-off is high installed electrical power and a refractory lining that wears with every campaign.

Channel induction furnaces work as a short-circuited transformer: a coil drives a primary winding on an iron core, and a loop of molten metal in a refractory channel acts as the single-turn secondary. They are highly efficient, roughly 10 to 15 percentage points above coreless units, but the channel must stay liquid, so they cannot melt from cold and serve mainly as large holding and superheating furnaces for copper, brass, and aluminum, with capacities reported from 8 to 120 tonnes and 300 to 2,500 kW per inductor.

Electric arc furnaces strike high-current arcs from graphite electrodes onto a scrap charge, melting it through intense radiant and conductive heat. They are the dominant route for recycling steel scrap, scaling from about 10 tonnes to the 400 tonne class, with modern tap-to-tap times under 60 minutes and twin-shell operations as fast as 35 to 40 minutes. Cupolas are vertical shaft furnaces charged from the top with alternating layers of iron, coke, and limestone flux; air blown through tuyeres burns the coke, and molten iron collects at the hearth for continuous tapping at 1 to 100 tonnes per hour. Crucible and reverberatory furnaces heat metal indirectly with a gas, oil, or resistance flame, and remain the simplest, lowest-capital option for non-ferrous and precious-metal melting.

Chapter 3 / 06

Heating Principles and Energy Sources

The heating principle determines melt rate, electrical efficiency, chemistry interaction, and capital cost. Four mechanisms dominate: electromagnetic induction, electric arc, fuel combustion, and electrical resistance. The table below compares their core engineering characteristics before the text examines each.

PrincipleHeat MechanismTypical EfficiencyEnergy per TonneChemistry Effect
Coreless inductionEddy currents in charge60 to 78%~520 to 650 kWh/t (steel)Low loss, electromagnetic stir
Channel inductionTransformer secondary loopup to ~95%~230 kWh/t (brass hold)Very low metal loss
Electric arcGraphite-electrode arc51 to 64%~550 to 700 kWh/t (steel)Oxidising, slag refining
CupolaCoke combustion40 to 50%100 to 150 kg coke/tCarbon and sulphur pickup

Electromagnetic induction passes alternating current through a copper coil, generating a magnetic field that induces eddy currents inside the metal charge; the charge heats itself through its own electrical resistance. Because the heat is generated in the metal rather than transferred across a surface, induction is fast, clean, and gives precise bath stirring. Coreless units typically achieve electrical efficiency in the 60 to 78 percent range and consume roughly 520 to 650 kWh per tonne of steel melted, depending on charge density and holding time. Frequency selection, from line frequency for the largest furnaces to 1,000 Hz or more for small crucibles, tunes penetration depth and stirring against charge piece size.

Electric arc heating draws a high-current arc, tens of kiloamperes, between graphite electrodes and the scrap. The arc plasma reaches several thousand degrees and melts the charge by radiation and conduction. EAFs are rated by transformer power, commonly 500 to 1,000 kVA per tonne and up to about 1,500 kVA per tonne in the most aggressive designs, and consume roughly 550 to 700 kWh per tonne of liquid steel, against a theoretical minimum near 300 kWh per tonne to melt scrap and a practical floor near 350 kWh per tonne in the best modern shops. The oxidising arc environment and a controlled slag also allow dephosphorisation and decarburisation, which is why the EAF is a refining furnace, not only a melter.

Fuel combustion burns coke in a cupola or natural gas and oil in crucible and reverberatory furnaces. A cupola consumes about 100 to 150 kg of coke per tonne of iron in modern hot-blast practice and needs roughly 800 to 900 cubic metres of air per tonne at a 10:1 iron-to-coke ratio. Combustion is cheaper per unit energy than electricity in many regions but couples the metal to the products of combustion, causing carbon and sulphur pickup in the cupola and oxidation losses in open-flame non-ferrous furnaces.

Electrical resistance heating uses silicon-carbide or molybdenum-disilicide elements, or an immersed resistor, to heat metal indirectly. It is slow but extremely controllable and clean, which suits aluminum holding furnaces, low-melting alloys, and laboratory melting where the priority is temperature precision rather than throughput.

A practical consequence of these differences is how each principle responds to scrap quality. Induction tolerates clean, dense, sized charge well but reacts poorly to oily, painted, or galvanized scrap, which fouls the bath and the fume system. The electric arc furnace, by contrast, is far more forgiving of mixed and contaminated scrap because its oxidising slag and high power burn off contaminants, which is one reason it became the recycling route of choice. Combustion furnaces sit in between, but open-flame contact raises oxidation loss on reactive metals such as aluminum and magnesium. Matching charge quality to principle is therefore as important as matching tonnage.

Chapter 4 / 06

Refractory Linings and Crucibles

The refractory lining is the consumable that contains the liquid metal, and its chemistry must match both the metal and the slag. A refractory material is conventionally defined as one whose refractoriness exceeds 1,580 degrees Celsius. The lining sees three attacks at once: thermal shock during heat-up and tapping, chemical attack from molten slag, and mechanical erosion from charging impact and electromagnetic stirring. Lining selection follows the acid-neutral-basic classification of slag chemistry, because an acid lining is rapidly destroyed by a basic slag and vice versa.

Acid silica linings are made from ramming mass of over 98 percent SiO2 with a small boric-acid sintering aid. They are the default for coreless iron and steel induction furnaces because they are inexpensive, easy to sinter into a hard face, and resistant to the acidic slags of grey-iron melting. Silica ramming mass is rated to roughly 1,600 degrees Celsius working temperature. Its limitation is poor resistance to basic slags and to thermal cycling at very high steel temperatures.

Neutral linings based on high-alumina (Al2O3) or alumina-spinel chemistry occupy the middle ground and suit steel, ductile iron, and operations where slag basicity varies. Basic magnesia (MgO) linings tolerate up to roughly 2,800 degrees Celsius and resist basic slags, which is why magnesia and magnesia-carbon brick line electric arc furnace hearths, steel ladles, and basic-melting induction furnaces. For aluminum and copper, low-cement alumina castables with non-wetting additives resist metal penetration and dross sticking.

Linings are installed in different physical forms: dry vibratory or rammed mass that is sintered in place, wet castable that is poured and cured, and pre-fired shaped brick. Small non-ferrous and precious-metal furnaces instead use a replaceable crucible, typically clay-graphite or silicon-carbide bonded, which is lifted out and swapped rather than relined. The table below maps common metals and slag conditions to a recommended lining for first-pass selection; always confirm against the maker's wear data and the actual slag analysis.

Metal / ServiceRecommended LiningApprox. LimitAvoid
Grey / ductile iron (coreless)Acid silica ramming mass~1,600 CBasic slag practice
Carbon and alloy steel (coreless)Alumina-spinel / magnesia~1,700 C+Plain silica
EAF hearth and ladleMagnesia-carbon brick~2,800 C (MgO)Silica, fireclay
Aluminum melting / holdingLow-cement alumina castable~1,400 to 1,600 CWettable silica
Copper, brass, bronzeSiC or clay-graphite crucible~1,300 to 1,500 CReactive linings
Precious metals (bench)Clay-graphite crucible~1,600 COversized crucible
Chapter 5 / 06

Key Specification Parameters

A melting furnace data sheet may list dozens of figures, but only a handful drive a purchase decision: rated capacity, installed power, frequency, melt rate, specific energy consumption, tapping temperature, and lining campaign life. Each is explained below, with the traps that catch first-time buyers.

Rated capacity is the working mass of liquid metal the furnace holds, quoted in kilograms or tonnes, and it is metal-specific: a crucible rated for 100 kg of aluminum holds roughly three times that mass of copper because of the density difference. Always confirm whether the figure is for steel, iron, aluminum, or copper, and whether it is the safe working capacity or the geometric brim-full volume. Installed power is the rated input of the power supply or transformer; for medium-frequency coreless iron furnaces it is commonly 600 to 1,000 kW per tonne of capacity, and for EAFs it is quoted as kVA per tonne, typically 500 to 1,000.

Frequency applies to induction furnaces and is decisive. Line frequency (50 or 60 Hz) suits the largest furnaces but needs a molten heel to start; medium frequency from roughly 150 to 1,000 Hz covers most foundry sizes and melts cold charges; higher frequencies of 1,000 to 3,000 Hz serve small crucibles. Lower frequency increases electromagnetic stirring, which helps alloying but, taken too far, erodes the lining and increases surface oxidation. Melt rate is the throughput in tonnes or kilograms per hour at full power from a defined start condition, and it must be read together with the start condition: cold-charge melt rates are far lower than hot-restart rates.

Specific energy consumption, in kWh per tonne for electric furnaces or kg of coke per tonne for a cupola, is the single most important operating figure because energy dominates lifetime cost. The table below lists representative figures by furnace type and metal; treat them as benchmarks, since real values shift with charge quality, holding time, and insulation.

Furnace / MetalSpecific EnergyTapping TempTypical Melt Rate
Coreless induction, steel~520 to 650 kWh/t~1,600 to 1,650 Cscales with kW/t
EAF, carbon steel~550 to 700 kWh/t~1,600 to 1,650 C<60 min tap-to-tap
Cupola, grey iron100 to 150 kg coke/t~1,400 to 1,500 C1 to 100 t/h
Channel induction, brass~230 kWh/t (hold)~1,000 to 1,100 C2 to 60 t/h
Crucible, aluminumvaries by fuel~700 to 750 C0.05 to 1 t batch

Tapping or pouring temperature is the bath temperature at which metal is removed; it sits above the melting point by the superheat needed to survive transfer and mould filling. Steel taps near 1,600 to 1,650 degrees Celsius, grey iron near 1,400 to 1,500, and aluminum near 700 to 750. Lining campaign life is the number of heats or weeks before relining, a recurring cost and planned outage that, for coreless iron furnaces, can range from a few weeks to a few months of campaigns. A furnace with cheap purchase price but short lining life and high energy use is often the most expensive over a five-year horizon.

Chapter 6 / 06

Selection Decision Factors

To convert the preceding chapters into a specific model, work through the ordered decision sequence below. Most melt-shop selection errors come not from a single wrong number but from settling furnace type before the downstream requirements are clear. These eight steps double as a fixed RFQ template.

  1. Metal and chemistry tolerance: Decide the metal (iron, steel, aluminum, copper, precious) and how tightly chemistry must be held. Induction gives the cleanest control; a cupola adds carbon and sulphur; an EAF can refine. This single answer eliminates most furnace families.
  2. Batch tonnage and throughput: Size the working capacity from peak hourly demand and pour size. Coreless induction suits 0.005 to 30 tonnes per batch; EAFs cover 10 to 420 tonnes; cupolas deliver 1 to 100 tonnes per hour continuously; crucibles handle small non-ferrous batches.
  3. Energy source and tariff: Compare local electricity price against coke or gas. Electric induction and arc give clean, controllable melting but high connected load; coke cupolas and gas crucibles can be cheaper per unit energy where fuel is inexpensive.
  4. Power supply and frequency: For induction, select line versus medium frequency from capacity and charge size, and confirm transformer or inverter rating in kW or kVA per tonne. Verify grid connection capacity and power-quality (flicker and harmonic) compliance, which matters acutely for EAFs.
  5. Refractory and crucible: Choose acid, neutral, or basic lining per Chapter 4 against metal and slag chemistry, or a replaceable crucible for small non-ferrous duty. Confirm expected campaign life and the cost and downtime of relining.
  6. Ancillary systems: Specify the closed-loop cooling-water plant, tilt hydraulics, dust extraction and fume hood, charging system, and temperature and chemistry instrumentation. These often cost as much as the furnace shell and determine real uptime.
  7. Safety and certification: Confirm compliance with IEC 60519 (Parts 1 and 3) for electroheat installations or NFPA 86 for fuel-fired and electric furnaces, plus PED 2014/68/EU and the EU Machinery Regulation 2023/1230 for ancillaries. Require a refractory leak or ground-leakage alarm that disconnects power before breakout.
  8. Total cost of ownership (TCO): Sum purchase price, installation, energy per tonne over the campaign, refractory relining, electrode or coke consumption, and downtime losses. An installation that saves on capital but consumes an extra 50 to 100 kWh per tonne usually loses that saving within the first year of production.

One dimension first-time buyers routinely undervalue is manufacturer serviceability: local field service, refractory and consumable supply, inverter or power-supply spares, and rapid response on a furnace breakout or coil failure. A melting furnace runs for fifteen to twenty-five years, and unplanned downtime on the melt deck stops the entire plant. Inductotherm, ABP Induction, Otto Junker, and Ajax TOCCO maintain global induction service networks, while large steel electric arc furnaces are supported by steel-plant engineering groups such as Danieli, SMS group, Primetals, and Tenova. For non-critical small non-ferrous melting, regional crucible-furnace builders are often the pragmatic choice on cost and lead time.

FAQ

What is the difference between a coreless and a channel induction furnace?

A coreless induction furnace surrounds a refractory crucible with a water-cooled copper coil; the entire charge sits inside the coil and is heated directly, so it can melt a cold solid charge from empty and is ideal for batch melting and alloy changes. A channel induction furnace works like a short-circuited transformer: a coil drives a primary winding wound around an iron core, and a loop of molten metal in a refractory channel forms the secondary. Because the channel must stay liquid, channel units cannot melt from cold and are used mainly as holding and superheating furnaces. Channel furnaces reach electrical efficiencies roughly 10 to 15 percentage points higher than coreless units, but they need a molten heel and suit lower-melting metals such as aluminum, copper, and brass.

How much power does a melting furnace need per tonne of capacity?

It depends on furnace type and metal. Medium-frequency coreless induction furnaces for iron and steel typically install roughly 600 to 1,000 kW per tonne of crucible capacity to reach a fast melt rate. Modern AC electric arc furnaces are rated at 500 to 1,000 kVA per tonne, and the most aggressive designs push toward 1,500 kVA per tonne of liquid steel. On energy consumed rather than installed, melting and superheating one tonne of steel needs about 550 to 700 kWh in a modern EAF, against a theoretical minimum near 300 kWh per tonne to melt the scrap and a practical floor near 350 kWh per tonne in the best shops. Aluminum melts at far lower energy because of its 660 degrees Celsius melting point, and brass holding in a channel furnace can run near 230 kWh per tonne.

What refractory lining should a melting furnace use?

The lining chemistry must match the slag and the metal. Acid silica ramming mass (over 98 percent SiO2) is the default for coreless iron and steel furnaces and is rated to about 1,600 degrees Celsius working temperature, because it is cheap, easy to sinter, and resists acid slags. Neutral high-alumina or alumina-spinel linings suit steel and ductile iron with basic or variable slags. Basic magnesia (MgO) linings, which tolerate up to roughly 2,800 degrees Celsius and basic slags, are standard for electric arc furnace hearths and steel ladles. Aluminum and copper furnaces favor low-cement alumina castables and non-wetting additives. Lining life is monitored by a leak-detection or ground-fault alarm that trips power before a metal breakout.

How do I choose between an induction furnace, an electric arc furnace, and a cupola?

Match the furnace to tonnage, metal, and scrap quality. Coreless induction furnaces dominate foundry melting from about 0.1 to 30 tonnes per batch because they give tight chemistry control, low melt loss, and clean operation; they suit iron, steel, and non-ferrous alloys. Electric arc furnaces are the choice above roughly 30 tonnes for bulk steel scrap recycling, with tap weights up to about 400 tonnes and tap-to-tap times under 60 minutes. Cupolas melt grey and ductile iron continuously at 1 to 100 tonnes per hour at low energy cost using coke, but they pick up carbon and sulphur and offer less chemistry control. Crucible furnaces fit small non-ferrous and precious-metal batches under about 0.5 tonne.

What frequency should a coreless induction melting furnace run at?

Frequency is chosen against crucible size and the metal. Line-frequency (50 or 60 Hz) furnaces suit very large capacities, typically above 10 to 25 tonnes, and need a molten heel to start. Medium-frequency furnaces from roughly 150 Hz to 1,000 Hz cover most foundry sizes, can melt a cold charge, and produce useful bath stirring for alloying. Higher frequencies of 1,000 to 3,000 Hz suit small crucibles down to a few kilograms, including precious-metal and laboratory units. Lower frequency increases electromagnetic stirring, which speeds alloy homogenization but, if excessive, can erode the lining and increase melt-surface turbulence and oxidation. Manufacturers tune frequency so penetration depth is matched to charge piece size.

What safety standards and certifications apply to melting furnaces?

Electrical melting installations are governed by the IEC 60519 series, Safety in installations for electroheating and electromagnetic processing: Part 1 covers general requirements and Part 3 covers induction and conduction heating and induction melting installations (BS EN 60519-3 in Europe). Fuel-fired and many electric furnaces in North America fall under NFPA 86, Standard for Ovens and Furnaces, which addresses fire and explosion hazards. Pressure vessel directives such as PED 2014/68/EU and the EU Machinery Regulation 2023/1230 apply to ancillary systems and machine safety. Practical requirements include a refractory leak or ground-leakage alarm that disconnects power before breakout, interlocked cooling-water flow, and emergency tilt and power-off controls.

What ongoing operating costs and serviceability factors matter when buying a melting furnace?

Total cost of ownership is dominated by energy, refractory, and downtime, not purchase price. Energy is the largest line item: an inefficient furnace that consumes an extra 50 to 100 kWh per tonne adds substantial annual cost at industrial tariffs. Refractory relining is a recurring cost and a planned outage; coreless iron linings may last from a few weeks to a few months of campaigns before push-out and re-ram. Cooling-water reliability, coil insulation life, power-supply (IGBT inverter) spares, and tilt-hydraulic maintenance determine uptime. Choose a maker with local field service, refractory supply, and inverter spares: Inductotherm, ABP Induction, Otto Junker, and Ajax TOCCO have global service networks, and large steel EAFs are supported by Danieli, SMS group, Primetals, and Tenova.

Ask SpecForge AI