Doc #106 — Electric Arc Furnace Operations: Regional Steelmaking from Scrap

---

RECOMMENDED ACTIONS

Phase 1 (Months 0–12): Planning and preservation

1. Inventory all EAF and induction furnace equipment in NZ — include NZ Steel Glenbrook (Doc #89), Pacific Steel Otahuhu,² all commercial foundries (Doc #93), university metallurgy departments (University of Auckland, University of Canterbury, University of Waikato), and any other facilities with electric melting capability. Record furnace type, capacity, power supply, transformer rating, condition, and consumable stocks.
2. Inventory all graphite electrodes in NZ outside Glenbrook — foundries, industrial suppliers, and any other holders. Every electrode is a strategic asset.
3. Classify foundry and steelmaking workers as critical-skills personnel. Prevent redeployment.
4. Secure all copper stocks relevant to induction furnace construction — copper tubing and bar suitable for coil fabrication. Coordinate with Doc #90 (copper recovery from scrap) and Doc #56 (wire and cable production).

Years 2–3: Site identification

5. Identify candidate sites for regional EAF/induction furnace facilities — criteria: proximity to grid power (minimum 1–5 MW available), proximity to scrap sources, existing heavy industrial infrastructure (foundations, overhead crane, water supply), and workforce accessibility. Leading candidates include: existing foundry sites (likely Auckland, Hamilton, Christchurch), former industrial sites with heavy power connections, and sites near scrap concentrations. Glenbrook handles NZ’s steel needs for years under reduced demand; regional EAF siting is a planning exercise, not an emergency.

Phase 2 (Years 1–3): Development

6. Commission engineering design for a prototype regional EAF — target 1–3 tonne capacity, using NZ-fabricated shell and refractories. The design should accommodate both graphite electrodes (from national stocks, rationed) and eventual Soderberg electrodes (Doc #89, Section 4.4).
7. Expand existing foundry induction furnace operations (Doc #93) to include steel melting — most foundry induction furnaces are set up for cast iron or non-ferrous alloys. Steel melting requires higher-temperature crucibles (magnesia or alumina-based rather than silica), higher power density, and different slag practice. This upgrade requires: sourcing or producing magnesia crucibles (NZ seawater magnesia program is not yet operational, so initial crucibles must be imported or fabricated from imported grain); developing slag practice and chemistry controls; installing a ladle and pouring system if not already present; and training operators in steel-specific temperature measurement and tap practice. These are achievable within existing foundry sites but represent meaningful development effort, not a trivial configuration change.
8. Begin Soderberg electrode paste development specifically for small-scale EAF use — the paste formulation (anthracite or carbonised wood, with coal tar or wood tar binder) is simpler than manufacturing graphite electrodes but requires experimentation to achieve adequate conductivity and mechanical strength.
9. Establish scrap preparation capability at regional sites — cutting, sorting, and sizing scrap for furnace charging. Scrap must be sorted to avoid tramp element contamination (Section 4 below).
10. Begin training EAF and induction furnace operators at existing facilities. Pair with Glenbrook knowledge capture (Doc #89) — Glenbrook operators understand arc furnace metallurgy that is broadly applicable to smaller-scale operations, though some adaptation is required: small furnaces have faster thermal cycles, less thermal inertia, and different electrode-to-bath geometry than Glenbrook’s large units.
11. Develop NZ-produced refractory linings for regional furnaces — smaller furnaces have less severe refractory requirements than Glenbrook’s EAFs, and NZ-produced fireclay and dolomite refractories (Doc #89, Section 5.4) may be adequate for lower-throughput operations.

Phase 3 (Years 3–7): First operations

12. Commission first regional EAF or induction furnace steelmaking facility. Target: producing basic carbon steel billets or ingots from sorted scrap, suitable for forging (Doc #92) or rolling into bar, rod, or simple sections.
13. Establish product capability for NZ’s most critical steel shortfall: wire rod and rebar. If Glenbrook has not adapted to produce wire rod (Doc #89, Section 7), regional facilities should prioritise rod and bar production from scrap using a small rolling mill or forging press.
14. Commission first submerged-arc furnace for calcium carbide production — this is a distinct project from steelmaking EAFs but uses similar technology (Section 8).
15. Establish second regional facility in a different part of the country to reduce transport distances and single-point-of-failure risk.

Phase 4+ (Years 7–15): Expansion

16. Expand regional facilities to 3–5 operating locations nationally — at minimum, one in the upper North Island (Auckland-Waikato), one in the lower North Island (Wellington-Manawatū), and one in the South Island (Canterbury).
17. Develop alloy steel capability — as operator skill matures, begin producing medium-carbon steel and basic alloy steels for tools, springs, and high-stress applications. This requires careful scrap selection and alloying additions (silicon from ferrosilicon produced in NZ submerged-arc furnaces; manganese if obtainable through trade).
18. Transition from graphite to Soderberg electrodes for all EAF operations, conserving any remaining graphite electrode stocks for Glenbrook or high-priority applications.
19. If trans-Tasman trade develops (Doc #138), prioritise import of graphite electrodes, ferroalloys, and power electronics (capacitor banks, thyristors) for induction furnace maintenance.

---

ECONOMIC JUSTIFICATION

The cost of not building regional steelmaking

If NZ relies solely on Glenbrook, then when Glenbrook’s graphite electrodes and critical refractories are exhausted — estimated at 1–3 years under reduced production (Doc #65, Section 4.5) — NZ loses all domestic steelmaking capacity. The only remaining steel supply is the existing stock of steel in the economy (finite and degrading) and whatever can be obtained through trade (uncertain and volume-limited). A nation that cannot produce steel cannot repair its hydro stations (Doc #65), maintain its water infrastructure (Doc #65), build agricultural equipment, construct ships (Doc #65), or sustain any industrial manufacturing. The cascading consequences of losing steelmaking are catastrophic.

Person-years to establish regional capability

Estimate: Establishing the first regional EAF or induction furnace steelmaking facility requires approximately 8–15 person-years of effort over 2–3 years, distributed across:

Activity	Person-years (est.)
Engineering design and planning	1–2
Furnace fabrication (shell, water cooling, electrode system or induction coil)	2–3
Refractory production and installation	1–2
Electrical installation (transformer, bus bars, power supply)	1–2
Rolling mill or forging capability (if included)	2–4
Operator training	1–2
Total per facility	8–15

This is a significant investment, but the alternative — total loss of steelmaking — renders it modest by comparison. A single 3-tonne EAF or induction furnace operating at two heats per day, 250 days per year, produces approximately 1,500 tonnes of steel annually. At 5–8 workers per shift, a facility producing 1,500 tonnes per year requires roughly 15–25 person-years of ongoing labour annually. This is far less efficient than Glenbrook (which produces several hundred thousand tonnes with ~1,500 workers), but the comparison is irrelevant — the comparison that matters is 1,500 tonnes per year versus zero.

Breakeven

There is no meaningful breakeven calculation because the alternative (no steel) is not an acceptable outcome. The question is not whether to build regional steelmaking but when and how.

---

1. WHY REGIONAL STEELMAKING

1.1 Glenbrook’s limits

Glenbrook (Doc #89) produces steel from virgin ironsand — a domestic ore supply that is effectively unlimited. The constraint on Glenbrook is not ore but consumables: graphite electrodes (estimated 1–3 years of supply under reduced production), refractory linings (longer runway but still finite), and irreplaceable equipment (EAF transformers, rolling mill rolls, process control systems). Glenbrook is managing depletion, not operating sustainably.

Regional EAF facilities face the same electrode constraint — but at a fundamentally different scale. A 3-tonne EAF melting clean scrap consumes approximately 2–5 kg of graphite electrode per tonne of steel,³ compared to Glenbrook’s 1.5–3 kg per tonne at far higher throughput. The absolute electrode consumption of a regional facility producing 1,500 tonnes per year is approximately 3–7.5 tonnes of electrodes annually — versus Glenbrook’s 1,000–2,000 tonnes per year. A small EAF can operate for years on an electrode stockpile that would last Glenbrook weeks.

Induction furnaces eliminate the electrode constraint entirely. They use copper coils and alternating current to heat the charge electromagnetically. No carbon electrode touches the metal. Their constraints are different (copper, capacitors, power electronics), but in a world where graphite electrodes are the binding constraint on steelmaking, induction furnaces offer a pathway forward.

1.2 The scrap opportunity

NZ’s accumulated steel stock — estimated at 15–30 million tonnes (Doc #90) — is a metal reserve that does not require mining, reduction, or the full Glenbrook process chain. Scrap steel is already metallic iron. Melting it requires approximately 350–500 kWh of electricity per tonne⁴ (compared to approximately 700–900 kWh per tonne for Glenbrook’s full ironsand-to-steel process, plus coal for the reduction kilns).⁵ The energy advantage of scrap is significant: NZ’s renewable grid can melt substantially more scrap per unit of electricity than it can process ironsand.

The metallurgical challenge is quality control. Scrap contains tramp elements — principally copper (from wiring in vehicles and appliances), tin (from coated steel), and zinc (from galvanised products) — that accumulate with each recycling pass and cannot be removed by standard steelmaking processes.⁶ Copper above approximately 0.2% causes hot shortness (cracking during hot rolling or forging). This is manageable through careful scrap sorting (Doc #90) and dilution with clean primary steel from Glenbrook while it operates, but it imposes real quality limits on scrap-based steel, particularly for products requiring hot working.

1.3 Geographic distribution

A single national steelworks at Glenbrook means all steel consumers depend on a single plant and the transport network connecting it to them. If the Cook Strait ferry service is disrupted (Doc #137), the entire South Island loses access to Glenbrook steel. Regional facilities distributed across both islands reduce this vulnerability and reduce the transport burden — scrap is collected and processed closer to where it is both sourced and consumed.

---

2. ELECTRIC ARC FURNACE TECHNOLOGY

2.1 How an EAF works

An electric arc furnace melts metal by passing very large electrical currents through graphite electrodes to create an electric arc — a sustained electrical discharge through the air gap between the electrode tips and the metallic charge. The arc generates temperatures of 3,000–3,500°C at the arc itself, heating the charge to its melting point (approximately 1,500–1,600°C for steel).⁷

A basic EAF consists of:

• A steel shell lined with refractory material, forming the furnace vessel. The shell is typically cylindrical with a dished bottom.
• A roof (also refractory-lined or water-cooled) that lifts or swings aside for charging.
• Graphite electrodes (typically three, in a three-phase AC configuration, or one in a DC configuration) that pass through holes in the roof. The electrodes are clamped in holders connected to a heavy electrical bus bar system.
• An electrode regulation system that raises and lowers the electrodes to maintain the arc as the charge melts and the surface level changes.
• A power supply — a step-down transformer that converts high-voltage grid supply (typically 11–33 kV) to the low voltage, very high current required by the arc (typically 100–400 volts at thousands of amperes for small furnaces).⁸
• A tapping mechanism — a spout or eccentric bottom tapping system to pour molten steel from the furnace into a ladle.
• A fume extraction system — EAF operation generates significant dust and fumes that must be captured for environmental and health reasons.

2.2 Scale considerations for NZ

Global mini-mill EAFs typically range from 30–150 tonnes per heat. These are far too large for NZ’s regional steelmaking needs and would consume graphite electrodes at rates NZ cannot sustain. The appropriate scale for regional NZ facilities is 1–10 tonnes per heat — the scale used by steel foundries and small specialty steelmakers worldwide.⁹

At this scale:

• Power requirement: A 3-tonne EAF requires approximately 1.5–3 MW of electrical power during the melt cycle (approximately 1–2 hours per heat).¹⁰ This is substantial — equivalent to approximately 500–1,000 NZ households at average consumption — but well within the capacity of NZ’s grid at any industrial connection point. For comparison, a medium-sized NZ dairy factory draws 1–3 MW during processing periods.
• Electrode consumption: Approximately 2–5 kg per tonne of steel, or 6–15 kg per heat. A single 500 mm diameter electrode column might last weeks to months of continuous operation at this scale.
• Refractory life: Small furnaces with basic refractory linings (dolomite rammed hearth, magnesia-carbon sidewalls) can achieve 100–300 heats per campaign before relining.¹¹ With NZ-produced dolomite and fireclay refractories, campaign life may be shorter — perhaps 50–150 heats. The practical consequence is that relining occurs more frequently: at 50 heats per campaign and two heats per day, relining is required every 3–4 weeks. Each reline requires 1–3 days of downtime and a skilled refractory crew. This reduces effective annual throughput by 10–25% compared to commercial refractory performance, and places a persistent demand on NZ’s refractory production program (Doc #89, Section 5.4).
• Throughput: Two heats per day (one melt-to-tap cycle in the morning, one in the afternoon) is a realistic operating pace for a small furnace with a trained crew. This produces approximately 6 tonnes of liquid steel per day, or roughly 1,500 tonnes per year at 250 operating days.

2.3 DC versus AC arc furnaces

Traditional EAFs use three-phase AC power with three electrodes. DC EAFs use a single electrode (cathode) with the return current passing through the furnace bottom (anode). DC furnaces have some advantages for small-scale operations:¹²

• Single electrode instead of three — reducing electrode consumption and simplifying electrode management
• Lower electrode consumption per tonne — typically 1–2 kg/tonne versus 2–5 kg/tonne for AC
• More stable arc — less electrical disturbance to the local grid (flicker), which matters for rural power connections
• Simpler power supply — a rectifier (AC to DC converter) is needed, but only one electrode regulation system

The disadvantage is that DC furnaces require a bottom electrode (anode) embedded in the furnace hearth, which is a more complex construction. DC power supply (rectification) also requires power electronics that may be difficult to maintain without imports.

Assessment for NZ: If suitable rectifier equipment is available (from industrial salvage or existing installations), DC EAFs are preferable for regional operations due to lower electrode consumption. If not, three-phase AC EAFs are the more straightforward construction.

2.4 What NZ can fabricate

A small EAF is, at its core, a steel vessel with a refractory lining, a roof with holes for electrodes, a transformer, and heavy electrical conductors. NZ can fabricate all structural components:

• Shell and roof: Plate steel from Glenbrook, cut and welded. NZ has extensive steel fabrication capability (Doc #89).
• Electrode holders and mast: Steel fabrication and machining. The electrode mast must allow vertical movement — a rack-and-pinion or hydraulic system is appropriate.
• Water cooling panels: For the upper sidewall and roof. Steel tubing welded into panels, connected to a cooling water circuit. Standard fabrication.
• Tapping system: A simple spout cut in the shell wall with a clay-and-sand plug that is broken open for each tap, or an eccentric bottom tapping (EBT) design for cleaner steel.
• Fume hood and ductwork: Sheet steel fabrication.

What NZ cannot fabricate (initially):

• The furnace transformer: This is the critical imported component. EAF transformers are specialised — they must deliver very high current at low voltage, withstand the electrical and mechanical stresses of arc operation, and include tap-changing capability to adjust power. NZ has transformer manufacturing capability for standard distribution transformers (Doc #69) and some power transformers, but EAF-duty transformers are a specialised category. However: at the 1–3 tonne furnace scale, transformer ratings are in the 1–5 MVA range, which is comparable to medium distribution transformers and well within NZ’s transformer rewinding capability.¹³ This is not the 50–100 MVA scale of Glenbrook’s EAF transformers.
• Electrode regulation controls: Modern EAFs use electronic controllers to maintain arc stability. For a small furnace, manual regulation (an operator observing the arc through a viewport or monitoring ammeter readings and adjusting electrode position with a manual control) is feasible — this is how EAFs were operated for decades before electronic control.¹⁴
• Power electronics for DC operation: Thyristor or diode rectifiers. These are imported components. If available from salvage or existing industrial installations, they can be repurposed.

---

3. INDUCTION FURNACE TECHNOLOGY

3.1 How induction melting works

An induction furnace heats metal by electromagnetic induction. A water-cooled copper coil surrounding a refractory crucible carries alternating current at a frequency of 50 Hz (mains frequency) to 10,000 Hz (medium frequency). This creates a rapidly alternating magnetic field that induces eddy currents in the metal charge. The electrical resistance of the metal to these currents generates heat — the metal heats itself from within, without any electrode contact or combustion.¹⁵

Two basic types:

• Coreless induction furnaces: The copper coil surrounds a cylindrical crucible. The entire charge is within the coil. This is the standard foundry melting furnace and the type most relevant to NZ’s regional steelmaking. Typical capacity: 50 kg to 30 tonnes.
• Channel induction furnaces: A loop of molten metal passes through a channel surrounded by a coil. Used for holding and superheating already-molten metal. Less relevant for scrap melting.

3.2 Advantages for NZ

Induction furnaces have several properties that make them strategically important for NZ’s post-Glenbrook future:

No graphite electrodes. This is the decisive advantage. Every tonne of steel melted in an induction furnace is a tonne that did not consume NZ’s dwindling graphite electrode stocks. In a world where electrodes are the binding constraint on steelmaking, induction furnaces represent the long-term future.

Cleaner metal. Induction-melted steel is not exposed to carbon from electrodes, reducing unintentional carbon pickup. This matters for producing low-carbon steel for drawing into wire (Doc #105) and other applications requiring controlled chemistry.

Lower noise and fume. Induction furnaces operate more quietly and produce less fume than arc furnaces. This matters for facilities in or near populated areas.

Flexible operation. An induction furnace can be started and stopped without damage — unlike an EAF where thermal cycling of refractories must be managed. This suits intermittent, campaign-style operation.

Existing NZ capability. NZ foundries already operate induction furnaces (Doc #93). The workforce and operational knowledge exist, though at foundry scale (typically 50 kg to 1 tonne), not steelmaking scale (3–10 tonnes).

3.3 Constraints and honest limitations

Copper. A coreless induction furnace coil for a 1-tonne steel melting furnace requires approximately 500–1,000 kg of copper tubing — thick-walled rectangular or round copper tube through which cooling water flows.¹⁶ NZ does not mine copper and has no smelting capability. Copper must come from NZ’s existing stock: recovered from electrical motors, cable, plumbing, and other scrap sources (Doc #90). Copper is one of NZ’s more constrained metals — every kg used for furnace coils is a kg not available for electrical wiring, motor repair, or other essential uses.

Power electronics. Mains-frequency (50 Hz) induction furnaces can operate directly from the grid through a standard transformer, but they are efficient only for larger furnaces (5+ tonnes) where the metal charge is large enough relative to the operating frequency to achieve good electromagnetic coupling.¹⁷ Smaller furnaces (0.5–3 tonnes) work better at medium frequency (500–3,000 Hz), which requires a frequency converter — either a motor-generator set (mechanically robust but inefficient) or a solid-state inverter (thyristors or IGBTs — imported electronics with finite life). NZ can build motor-generator sets; it cannot manufacture power semiconductor devices.

Refractory crucibles. Induction furnace crucibles for steel melting must withstand temperatures exceeding 1,600°C and are typically made from rammed magnesia or alumina. These are consumed — a crucible may last 50–200 heats before replacement.¹⁸ NZ’s refractory development program (Doc #97, Section 5) applies here, with the smaller crucible scale making NZ-produced refractories more feasible than for Glenbrook’s large EAF linings.

No refining capability. An induction furnace is a melting device, not a refining device. It cannot reduce carbon or phosphorus from the melt the way an EAF or oxygen converter can. The chemistry of the output steel is essentially the chemistry of the input scrap, minus some oxidation losses. This means scrap must be carefully sorted to the desired chemistry before charging. For critical applications, a secondary refining step (ladle treatment with lime and flux) is needed.

Scale ceiling. Practical induction furnace size for NZ is limited by copper availability (for coils), power supply capacity, and the frequency converter constraint. Furnaces larger than about 5 tonnes require mains-frequency operation and proportionally more copper. A realistic national fleet of induction furnaces might comprise 5–15 units of 0.5–3 tonne capacity, distributed across regional facilities.

3.4 What NZ can build

The coil: Requires copper tubing. If suitable tubing is recovered from existing NZ stocks (copper plumbing tube, refrigeration tube, or recovered from end-of-life equipment), coils can be wound and brazed locally. This requires: a winding mandrel (steel former machined to the crucible diameter); silver or phosphor-copper brazing alloy for the inter-turn joints (NZ has brazing consumable stocks, though these are finite); a propane or oxy-acetylene torch; and the electromagnetic design calculations to determine turn count, coil geometry, and water flow rates. The design work is within the scope of an electrical engineering graduate with access to induction heating reference texts.¹⁹ Coil fabrication is therefore feasible in NZ but not trivial — it requires specific materials (brazing alloy) and engineering competence, not just metalworking skill.

The crucible: Rammed from NZ-produced refractory material into a steel form. The crucible is formed in place — dry refractory grain is packed around a former, then sintered during the first few heats.

The power supply: For mains-frequency operation, a standard step-down transformer (within NZ’s manufacturing capability) is sufficient. For medium-frequency operation, a motor-generator set can be built from a standard electric motor driving a higher-frequency generator — both within NZ’s motor rewinding and electrical engineering capability (Doc #95), though with efficiency losses compared to solid-state inverters. A motor-generator set typically achieves overall electrical efficiency of 80–88%, versus 93–97% for a solid-state IGBT inverter; this increases energy cost per heat by roughly 10–20%. There is also a practical frequency ceiling: motor-generator sets for furnace service are typically limited to 500–1,000 Hz, whereas solid-state inverters can reach 3,000 Hz — the lower frequency means somewhat larger minimum furnace size for efficient coupling.²⁰

The tilting mechanism: The furnace must tilt to pour molten metal into a ladle. Hydraulic or mechanical tilting frames are standard steel fabrication.

The cooling system: Water-cooled copper coils require a recirculating water system with heat dissipation (cooling tower or simple radiator). Standard plumbing and pump installation.

---

4. CONSUMABLE MANAGEMENT

4.1 Graphite electrodes: The binding constraint for EAF operations

Graphite electrode supply for regional EAFs comes from three sources:

1. National stockpile — electrodes requisitioned from NZ Steel, foundries, and industrial suppliers (Doc #1). These must be allocated by a national authority that balances Glenbrook’s needs against regional facility needs. Glenbrook, while operating, has higher priority — it produces more steel per tonne of electrode consumed.

2. Soderberg (self-baking) electrodes — if the Soderberg development program (Doc #89, Section 4.4) succeeds, NZ can produce electrode paste from carbonised wood or NZ coal mixed with wood tar or coal tar binder. Soderberg electrodes are lower quality than graphite — lower electrical conductivity means higher resistive heating losses in the electrode column, higher consumption rate, and less stable arc characteristics.²¹ But they can be produced from NZ materials indefinitely. The performance gap is real but acceptable for small-scale operations where electrode consumption per heat is modest.

3. Carbon electrodes from charcoal — a more speculative pathway. Carbon electrodes (as distinct from graphite electrodes) can be produced by pressing and baking a mixture of calcined petroleum coke or high-quality charcoal (Doc #102) with a tar pitch binder at approximately 800–1,000°C.²² These are not graphitised (which requires 2,500–3,000°C — an extremely energy-intensive process requiring specialised Acheson furnaces). Ungraphitised carbon electrodes have much higher electrical resistivity than graphite (roughly 10–50x), which limits their suitability for high-power EAF operation. For small furnaces operating at lower power density, they may be functional — but with significantly higher electrode consumption and less efficient melting. This requires experimental validation.

Strategy: During Phase 2–3, regional EAFs use graphite electrodes from the national stockpile, rationed carefully. As Soderberg and carbon electrode production develops (Phase 3–4), operations transition away from graphite. Induction furnaces are prioritised for expansion because they bypass the electrode constraint entirely.

4.2 Refractories

Regional furnaces have less severe refractory requirements than Glenbrook’s EAFs for two reasons: smaller size means lower thermal and mechanical stress, and lower throughput means less chemical erosion per unit time.

NZ-produced refractories suitable for regional EAFs and induction furnaces:

• Rammed dolomite hearth (EAF): Calcined NZ dolomite, rammed and sintered in place. Adequate for basic EAF bottoms. Campaign life estimated at 50–200 heats.²³
• Fireclay sidewall brick (EAF): From NZ fireclay deposits. Suitable for the upper sidewall where temperatures are lower. Not suitable for the slag line (the zone of maximum chemical attack).
• Magnesia crucible (induction furnace): From seawater magnesia production (Doc #98, Section 5.3) once developed. In the interim, silica-based crucibles from NZ silica sand are adequate for cast iron melting but not for steel (silica reacts with steel at steelmaking temperatures, contaminating the melt with silicon and eroding the crucible). Alumina crucibles from imported stocks or NZ bauxite substitutes are a middle option.
• Seawater magnesia development is the key enabling technology for long-term refractory self-sufficiency. The Dow process for precipitating magnesia from seawater using lime is well-established chemistry.²⁴ A pilot plant should be a Phase 2–3 priority (Doc #112).

4.3 Fluxes and slag practice

EAF steelmaking requires lime (calcium oxide) and sometimes fluorspar (calcium fluoride) to form a slag that absorbs impurities (phosphorus, sulfur, inclusions) from the steel. NZ has domestic limestone deposits, and lime is produced at several NZ kilns (Doc #97, Section 3.4). Fluorspar is not available in NZ — lime-only slag practice is adequate for basic steelmaking, though with somewhat less effective sulfur removal.²⁵

For regional operations melting sorted scrap, slag practice is simpler than for Glenbrook’s ironsand-based process (which produces titanium-rich slag requiring specialised management). A basic lime-based slag added to the EAF after melting — approximately 30–50 kg of lime per tonne of steel — is sufficient for most applications.

4.4 Deoxidisers and alloying additions

Liquid steel contains dissolved oxygen that must be removed before casting (deoxidation). Standard deoxidisers are ferrosilicon (FeSi) and ferromanganese (FeMn) — both currently imported.

Ferrosilicon can be produced in NZ using a submerged-arc furnace (Section 8) from NZ silica sand and charcoal (Doc #102). This is a significant development project but uses entirely domestic materials and grid electricity.

Ferromanganese requires manganese ore. NZ has small manganese deposits in Northland (including occurrences near Waihopo/Kaikohe) and the East Cape region that have not been commercially exploited.²⁶ Whether these deposits are economically viable at the small scale needed is uncertain — they would require geological assessment, extraction planning, and ore beneficiation capability before any ferromanganese production could proceed. Alternatively, manganese-containing scrap (such as rail steel and some structural steels, which typically contain 0.5–1.5% manganese) can be used as the charge — melting manganese-rich scrap inherently provides manganese to the melt.

Aluminium for deoxidation: small quantities of aluminium (1–2 kg per tonne) are used as a powerful deoxidiser. NZ’s aluminium supply depends on the Tiwai Point smelter status (Doc #109, Section 6.2). If aluminium is not available, silicon and manganese deoxidation is adequate for most structural applications, though steel cleanliness (inclusion content) will be inferior.

---

5. PRODUCTS: WHAT REGIONAL FACILITIES SHOULD MAKE

5.1 Priority products

The products that NZ most urgently needs but Glenbrook may not produce include:

Wire rod (5.5–12 mm diameter round bar): The starting material for wire drawing (Doc #105). Wire rod is the highest-priority product because it enables fencing wire, nails, barbed wire, springs, and wire rope — all critical for agricultural and industrial recovery. Wire rod is produced by hot rolling billets through a series of reducing passes. A small rod mill capable of rolling 5.5–12 mm rod from 75–100 mm square billets is within NZ’s fabrication capability, though it is a significant engineering project.

Reinforcing bar (rebar, 10–25 mm diameter deformed bar): Essential for concrete construction (Doc #97). Rebar is simpler to produce than wire rod (larger diameter, less reduction required) and is the standard product of scrap-based mini-mills worldwide.

Round and flat bar stock: For machining (Doc #91), forging (Doc #91), and general fabrication. Bar stock in standard sizes (10–50 mm round, 10–50 mm flat) is consumed by every engineering workshop in the country.

Billets and ingots: If rolling capability is not immediately available, the simplest product is cast billets (square or round section, 75–200 mm) or ingots. These can be forged into shapes by blacksmiths (Doc #92) or stockpiled for later rolling.

5.2 Casting versus rolling

The simpler production pathway is to cast steel into moulds — producing ingots, billets, or even near-net-shape castings (valve bodies, bearing housings, structural nodes) using the foundry techniques described in Doc #93. This requires no rolling mill.

Rolling produces longer, continuous products (rod, bar, sections) at higher throughput but requires a rolling mill — rolls, bearings, a drive system, and reheating capability. A small rolling mill for rod and bar production is a substantial but feasible project:

• Rolls: Can be cast from high-carbon steel or grey cast iron in NZ foundries (Doc #93). Roll surface quality will be inferior to imported precision-ground rolls, producing rougher surface finish on the rolled product. For rebar and fencing wire, this is acceptable.
• Roll bearings: Heavy-duty bearings from NZ stocks or trade. Alternatively, plain bronze bearings (bushing-type) can be cast from NZ-available copper alloys — lower speed and shorter life than roller bearings but functional for intermittent operation.
• Drive system: Electric motor with gear reduction. Within NZ’s electrical and mechanical capability.
• Reheating furnace: Gas or electric furnace to heat billets to rolling temperature (~1,100–1,200°C). Natural gas (Taranaki), wood gas (Doc #56), or direct electric resistance heating are all options.

Assessment: Casting is achievable in Phase 3. Rolling requires Phase 3–4 development. Both should be pursued — casting first for immediate needs, rolling as capability matures.

5.3 Product quality

Steel produced from scrap in small regional furnaces will be of lower and less consistent quality than Glenbrook’s primary steel. Specifically:

• Tramp elements: Copper, tin, and other contaminants from imperfectly sorted scrap. Can be managed through sorting (Doc #89) but not eliminated.
• Inclusion content: Higher than clean primary steel, due to less sophisticated refining and deoxidation. Reduces fatigue life and ductility.
• Chemical consistency: Variable from heat to heat unless rigorous scrap sorting and furnace practice are maintained.
• Surface quality of rolled products: Rougher than products from modern rolling mills.

For most recovery applications — rebar, fencing wire, general structural bar, castings — this quality level is adequate. For critical applications (springs, high-stress fasteners, wire rope), the best-quality heats should be selected and tested before use. Quality control through simple mechanical testing (bend tests, tensile tests using a basic testing machine) is essential and within NZ’s existing materials testing capability (universities, HERA, and larger engineering firms have testing equipment).²⁷

---

6. SITING AND INFRASTRUCTURE

6.1 Power supply

A 3-tonne EAF drawing 2–3 MW is a significant but not extraordinary load on NZ’s grid. Industrial power connections at this level exist at many NZ industrial sites. The constraint is not total grid capacity (NZ generates approximately 42,000–44,000 GWh annually²⁸) but local connection capacity — the site must have a sufficiently rated transformer and switchgear to deliver 2–3 MW at the voltages required.

Ideal sites have existing heavy industrial connections: former freezing works, dairy factories, engineering workshops, or industrial zones. Auckland, Hamilton, Palmerston North, Christchurch, and Dunedin all have industrial areas with suitable power supply.

Grid impact: EAF operation causes voltage fluctuations (flicker) on the local network due to the variable nature of the arc. This is more pronounced for AC furnaces. Mitigation includes operating on strong grid connection points (near transmission substations rather than at the end of long distribution feeders) and, if available, static VAR compensators (SVCs) from industrial salvage.²⁹ At the 1–3 MW scale, the grid impact is modest — comparable to large industrial motor starts, which NZ’s grid already manages.

6.2 Water supply

Both EAF and induction furnaces require cooling water — for electrode holders and shell panels (EAF) or coils (induction furnace). Water consumption is primarily recirculating (closed loop with a cooling tower), not consumptive. A facility needs access to approximately 10–20 m³/hour of makeup water to compensate for evaporation and blowdown.³⁰ This is easily available at any NZ site with a municipal water connection or a bore/stream supply.

6.3 Candidate locations

Upper North Island — Auckland/Waikato region: - Closest to Glenbrook (for workforce exchange and technology transfer) and to Pacific Steel Otahuhu - Largest concentration of engineering workshops and industrial infrastructure - Best access to North Island scrap stocks (Auckland is NZ’s largest scrap source) - Existing foundry operations in the region - Strong grid connections via Transpower and Vector networks

Lower North Island — Palmerston North/Hutt Valley: - Manawatū has several existing engineering firms with heavy industrial capability, including firms servicing the agricultural and dairy sectors - Lower Hutt (Seaview/Gracefield industrial area) has a history of industrial manufacturing and retains heavy industrial zoning with suitable power infrastructure - Serves the lower North Island and Wellington region without dependence on Auckland supply chains - Transpower’s grid connection at Haywards substation provides strong grid access for heavy industrial loads

South Island — Christchurch/Canterbury: - South Island’s largest industrial base - Existing engineering workshops and some foundry capability - Reduces South Island dependence on North Island steel supply (which requires Cook Strait shipping) - Proximity to scrap sources from Canterbury farming region and Christchurch’s building stock

6.4 Environmental and safety considerations

EAF steelmaking produces significant noise (arc noise during melting, 90–110 dB at the furnace shell), dust (metallic fume from the arc and from oxidising scrap), and slag (which must be managed as waste or used in road construction). These are manageable through:

• Fume extraction: An enclosed roof design with ducting to a baghouse (fabric filter dust collector). Baghouse systems are within NZ’s fabrication capability — the critical component is the filter fabric, which is currently imported but could potentially be produced from NZ wool or synthetic fabric from existing stocks.
• Noise management: EAF noise can be managed through building enclosure and siting away from residential areas.
• Slag disposal: EAF slag from scrap melting is primarily iron oxide, calcium silicates, and tramp elements. It can be used as aggregate for road construction or stored. It is not hazardous but should not be placed where it could leach into waterways.
• Safety: Molten steel at 1,600°C is inherently dangerous. Furnace operations require strict safety protocols, personal protective equipment (heat-resistant clothing, face shields, insulated gloves), and trained operators. NZ’s existing foundry and heavy engineering workforce understands these risks.

---

7. WORKFORCE AND TRAINING

7.1 Skill requirements

Operating a small EAF or induction furnace requires a team with the following skills:

• Furnace operator (melter): Understands charge calculation (what scrap to add to achieve target chemistry), melting procedure, temperature measurement, slag management, and tapping. This is the most skilled role.
• Crane/charging operator: Manages scrap charging into the furnace using an overhead crane or charging system.
• Ladle and casting operator: Receives molten steel from the furnace, manages ladle treatment (flux addition, deoxidation), and pours into moulds or a continuous casting system.
• Electrician: Maintains power supply, electrode regulation, and induction furnace electronics.
• Refractory worker: Installs and repairs furnace linings.
• Scrap preparer: Sorts, sizes, and prepares scrap charges (Doc #90).

A minimum crew for a small EAF or induction furnace facility is approximately 5–8 people per shift, with a total workforce of 15–25 for two-shift operation with maintenance and support staff.

7.2 Training pathway

Phase 2 (planning period): - Recruit from existing NZ foundry workers (Doc #93), Glenbrook operators being cross-trained (Doc #89), and skilled trades (welders, fitters, electricians) willing to retrain. - Conduct training at existing foundry sites using foundry induction furnaces as training equipment. - If Glenbrook is still operating, arrange for trainee placements to observe and participate in EAF operations. - Adapt and print relevant sections of standard metallurgical textbooks — The Making, Shaping and Treating of Steel (AISE/AIST), Electric Furnace Steelmaking (various authors), foundry operation manuals. University of Auckland and University of Canterbury metallurgy department libraries should hold copies.³¹

Phase 3 (commissioning): - First heats at a new facility should be conducted with the most experienced available operators, progressing to full crew operation as confidence builds. - Establish systematic quality control from the first heat — every heat produces a sample for testing. - Document all operational experience — what worked, what failed, what was learned. This is NZ’s nascent steelmaking knowledge base, and it must be recorded.

---

8. SUBMERGED-ARC FURNACE PRODUCTS

Phase: 3–5 (calcium carbide and ferrosilicon Phase 3–4; silicon metal Phase 5+) | Feasibility: [B] Feasible for calcium carbide and ferrosilicon; [C] Speculative for silicon metal

8.1 What a submerged-arc furnace is

A submerged-arc furnace (SAF) is a type of electric arc furnace where the electrodes are buried in the charge material rather than arcing across an open gap. The arc forms within the charge, generating heat internally. SAFs are used industrially to produce a range of electrochemical products from raw materials by carbothermic reduction — reacting an oxide ore with carbon at high temperature to produce the metal or a metal-carbon compound.³²

SAFs use the same basic equipment as EAFs — graphite or Soderberg electrodes, a refractory-lined vessel, a power supply — but the operating conditions and products are different. Critically for NZ, SAFs can use Soderberg electrodes more readily than open-arc EAFs, because the electrode column is buried in the charge and the arc conditions are less demanding.³³

8.2 Calcium carbide

Calcium carbide (CaC₂) is produced by reacting lime (CaO) with carbon (coke or charcoal) in a submerged-arc furnace at approximately 2,000°C:

CaO + 3C → CaC₂ + CO

NZ has both raw materials: lime from domestic limestone (produced at NZ kilns), and charcoal from NZ forestry (Doc #97). The energy source is grid electricity.

Calcium carbide matters for recovery because it reacts with water to produce acetylene gas:

CaC₂ + 2H₂O → C₂H₂ + Ca(OH)₂

Acetylene is essential for oxy-acetylene cutting and welding (Doc #94), which is the primary thermal cutting process available to NZ workshops. Without calcium carbide, oxy-acetylene capability ceases when existing acetylene stocks are exhausted. Current acetylene in NZ is produced from calcium carbide by industrial gas suppliers (BOC/Linde), using imported calcium carbide.³⁴ Domestic carbide production closes this import dependency.

Scale: NZ’s acetylene consumption is modest by global standards. A small SAF producing 2,000–5,000 tonnes of calcium carbide per year — a furnace drawing approximately 2–5 MW — would likely cover NZ’s recovery-era demand for welding and cutting gas.³⁵

8.3 Ferrosilicon

Ferrosilicon (FeSi) is an alloy of iron and silicon produced by carbothermic reduction of silica (SiO₂) with iron and carbon in a submerged-arc furnace:

SiO₂ + 2C + Fe → FeSi + 2CO

NZ has silica sand (from several deposits, including Parengarenga Harbour in Northland — one of the world’s highest-purity silica sand deposits)³⁶, iron (from Glenbrook steel or scrap), and carbon (charcoal from Doc #102). Grid electricity provides the energy.

Ferrosilicon is used as: - Steel deoxidiser: Added to molten steel to remove dissolved oxygen (Section 4.4) - Inoculant in cast iron: Controls graphite formation in iron castings (Doc #93) - Silicon source for other chemical processes

Without ferrosilicon, NZ’s steelmaking and foundry operations are severely compromised. Domestic ferrosilicon production using a small SAF (1–3 MW) producing a few hundred tonnes per year would satisfy NZ’s recovery-era needs. The process is among the simpler SAF applications, but still requires: a purpose-built refractory-lined furnace vessel with Soderberg or graphite electrodes; a power supply rated 1–3 MW at low secondary voltage; consistent silica sand, iron scrap, and charcoal feedstock; slag management and tapping infrastructure; and trained operators capable of reading furnace behaviour and managing tap chemistry. This is a Phase 3 engineering project, not a workshop-scale operation.³⁷

8.4 Silicon metal

At higher purity and with different operating conditions, the same SAF process can produce metallurgical-grade silicon metal (98–99% Si) rather than ferrosilicon. Silicon metal is the starting material for:

• Silicone sealants and lubricants (with substantial further chemistry)
• Solar-grade silicon (with extensive further purification — the Siemens process, requiring trichlorosilane chemistry, or fluidised bed reactor deposition)
• Chemical applications

Silicon metal production from silica sand is theoretically feasible in NZ but requires higher furnace temperatures, more careful process control, and higher-purity raw materials than ferrosilicon. This is a Phase 5+ development target, not a near-term goal.³⁸

---

9. STAGED DEVELOPMENT PLAN

9.1 Stage 1 — Foundry-scale melting (Phase 2–3, Years 1–5)

Expand existing NZ foundry induction furnace operations (Doc #93) from casting-focused melting to include basic steel melting. This means: - Upgrading crucible refractories from silica (adequate for cast iron) to magnesia or alumina (required for steel) - Increasing charge weight to 0.5–1 tonne per heat - Developing slag practice for steel (lime-based) - Establishing quality control (chemical analysis of each heat by optical emission spectrometry — NZ materials testing labs have this equipment)

Output: Steel ingots and castings. Perhaps 200–500 tonnes per year per foundry, with 3–5 foundries operational.

9.2 Stage 2 — First dedicated EAF (Phase 3–4, Years 3–10)

Build and commission the first purpose-built small EAF (1–3 tonnes per heat) at a selected regional site. This facility should include: - The EAF itself (NZ-fabricated shell, imported or NZ-produced refractories, graphite or Soderberg electrodes) - A ladle treatment station (lime and deoxidiser addition) - An ingot/billet casting station (moulds cast from NZ iron — Doc #93) - Eventually, a small bar/rod rolling mill

Output: 1,000–2,000 tonnes per year of steel billets, ingots, and (once rolling is available) bar and rod products.

9.3 Stage 3 — Network of regional facilities (Phase 4–5, Years 7–20)

Replicate the Stage 2 facility at 2–4 additional locations. Each facility develops product specialisation based on local needs: - Upper North Island: wire rod and rebar (serving agricultural and construction demand) - Lower North Island: bar and structural products - South Island: bar, castings, and (if Tiwai Point aluminium is available) non-ferrous products

Output: 5,000–30,000 tonnes per year nationally across all facilities.

9.4 Stage 4 — SAF products (Phase 3–5, Years 3–20)

Parallel to steelmaking, establish submerged-arc furnace production of: - Calcium carbide (Phase 3): enables domestic acetylene production - Ferrosilicon (Phase 3–4): enables steel deoxidation and foundry inoculation - Silicon metal (Phase 5+): prerequisite for advanced materials

Output: Calcium carbide: 2,000–5,000 tonnes/year. Ferrosilicon: 200–1,000 tonnes/year.

---

10. CRITICAL UNCERTAINTIES

Uncertainty	Impact if wrong	Resolution pathway
Graphite electrode stocks in NZ (outside Glenbrook)	Determines how many EAF heats are possible before transition to Soderberg or induction	National inventory — Phase 1 priority (Doc #1, Doc #8)
Soderberg electrode quality from NZ materials	If NZ coal and wood tar produce inadequate paste, EAFs become unviable after graphite depletion	Development trials — Phase 2
Carbon electrode feasibility for small EAFs	If ungraphitised carbon electrodes work in small furnaces, electrode supply broadens significantly	Experimental program — Phase 2–3
Copper availability for induction furnace coils	If insufficient copper is recoverable from scrap, induction furnace construction is constrained	Copper audit as part of national scrap census (Doc #8)
Power electronics availability for medium-frequency induction	If thyristors and IGBTs are not available, only mains-frequency (large) induction furnaces are feasible	Inventory of NZ industrial electronics and salvageable power equipment
Seawater magnesia production timeline	Determines when NZ-produced refractories are available for furnace linings	Pilot plant development — Phase 2–3
Scrap quality after sorting	If tramp elements are higher than expected, product quality is lower and application range narrower	Scrap sorting programme (Doc #156) and heat-by-heat testing
NZ foundry workforce availability	If existing foundry workers are too few or cannot be retained, training lead time increases	Skills census (Doc #8) — Phase 1
Rolling mill feasibility at small scale	If rolls, bearings, or drives cannot be sourced or fabricated, product range is limited to castings and ingots	Engineering assessment — Phase 2
Local SAF operation for calcium carbide and ferrosilicon	If NZ raw materials (charcoal, lime, silica) produce satisfactory products, multiple import dependencies are eliminated	Trial production — Phase 3

---

CROSS-REFERENCES

• Doc #1 — National Emergency Stockpile Strategy (electrode and copper requisition)
• Doc #8 — National Asset and Skills Census (foundry and steelmaking workforce identification)
• Doc #33 — Tires (steel wheels as tire alternative for slow transport)
• Doc #70 — Copper Wire Production (copper recovery for induction coils)
• Doc #56 — Wood Gasification (reheating furnace fuel)
• Doc #65 — Hydroelectric Station Maintenance (steel castings and forgings for hydro equipment)
• Doc #67 — Transpower Grid Operations (power supply for EAF facilities)
• Doc #69 — Transformer Maintenance and Rewinding (furnace transformer construction)
• Doc #74 — Pastoral Farming Under Nuclear Winter (fencing wire demand driver)
• Doc #89 — NZ Steel Glenbrook (primary steelmaking, electrode and refractory technology, Soderberg development)
• Doc #90 — Scrap Metal as Resource (feedstock for regional furnaces)
• Doc #91 — Machine Shop Operations (fabrication of furnace components, consumer of bar stock)
• Doc #92 — Blacksmithing and Forge Work (consumer of ingots and bar, direct scrap reworking)
• Doc #93 — Foundry and Metal Casting (existing induction furnace base, casting capability)
• Doc #94 — Welding Consumables (calcium carbide for acetylene, consumer of steel products)
• Doc #95 — Motor Rewinding (motor-generator sets for induction furnaces)
• Doc #97 — Cement and Concrete (rebar demand driver)
• Doc #102 — Charcoal Production (carbon for SAF processes, charcoal for Soderberg paste)
• Doc #105 — Fencing Wire, Nails, and Wire Drawing (wire rod demand driver)
• Doc #113 — Sulfuric Acid (potential shared industrial chemistry infrastructure)
• Doc #138 — Sailing Vessel Design (trade route for electrode and ferroalloy imports)
• Doc #151 — Trans-Tasman Relations and Trade (Australian electrode and ferroalloy imports)
• Doc #157 — Trade Training (steelworker and foundry training pipeline)
• Doc #160 — Heritage Skills Preservation (traditional metallurgical knowledge; iwi partnership frameworks for regional facilities, §4.5–4.7)

---

---

1. The 20,000–80,000 tonne range is a planning estimate, not a modelled figure. The lower bound assumes 5 facilities producing ~1,500 tonnes/year each (2 heats/day, 250 days/year, 3-tonne EAF or induction furnace), with roughly a third of those offline at any time for relining or maintenance. The upper bound assumes 10–15 facilities including some larger units (5–10 tonne class) operating more reliably as consumable constraints ease. Actual output will depend critically on electrode and copper availability, workforce depth, and grid reliability. This estimate should be revised as Phase 2 census data and facility commissioning experience becomes available.↩︎

2. Pacific Steel, located at Otahuhu, Auckland, is NZ’s primary wire and reinforcing steel producer. Pacific Steel operates an electric arc furnace that melts scrap steel to produce wire rod, reinforcing bar, and other long products. It is owned by Fletcher Building Ltd. Pacific Steel’s existing EAF operation makes it the most directly relevant precedent for the regional EAF capability described in this document. See: Fletcher Building corporate information; Pacific Steel product data. https://www.fletcherbuilding.com/↩︎

3. Graphite electrode consumption in scrap-based EAF steelmaking is typically 1.5–5 kg per tonne of liquid steel, depending on furnace design (AC vs DC), electrode quality, power level, and operating practice. Scrap charges melt faster and with less arc energy than direct-reduced iron (DRI), resulting in lower electrode consumption per tonne compared to Glenbrook’s DRI-based process. See: Fruehan, R.J. (ed.), The Making, Shaping and Treating of Steel — Steelmaking and Refining Volume, AISE, 1998; Madias, J., “Electric Furnace Steelmaking,” in Treatise on Process Metallurgy, Elsevier, 2014.↩︎

4. Electrical energy consumption for melting scrap steel in a modern EAF is approximately 350–500 kWh per tonne of liquid steel. This figure includes melting, superheating, and holding energy. Smaller furnaces tend toward the higher end of the range due to higher proportional heat losses. See: International Energy Agency, “Energy Technology Perspectives — Iron and Steel”; AIST (Association for Iron & Steel Technology) technical publications.↩︎

5. Glenbrook’s total energy consumption (electricity plus coal) for the full ironsand-to-steel process is significantly higher than scrap melting alone. The electrical component for the EAF melting step is approximately 700–900 kWh per tonne, with additional coal energy for the reduction kilns. Total primary energy per tonne of steel at Glenbrook is estimated at 15–20 GJ, versus approximately 5–8 GJ for scrap-based EAF steelmaking. Based on general steelmaking energy data and NZ Steel’s reported consumption figures. See: World Steel Association energy statistics; MBIE NZ Energy Data Tables.↩︎

6. The tramp element problem in scrap recycling — particularly copper accumulation — is well-documented in steelmaking metallurgy. Copper cannot be removed by any standard steelmaking process (oxidation, slag treatment, vacuum degassing) because it is thermodynamically more noble than iron. See: Savov, L. et al., “Problems of Recycling in Steelmaking,” Journal of the University of Chemical Technology and Metallurgy, 38(4), 2003; Daehn, K.E. et al., “Transforming the Steel Cycle Through Improved Copper Management,” Nature Sustainability, 2022.↩︎

7. EAF arc temperatures and melting fundamentals: The arc temperature exceeds 3,000°C at the electrode tip, with the bath reaching 1,550–1,700°C depending on the steel grade. See any standard steelmaking text, e.g., Ghosh, A. and Chatterjee, A., Ironmaking and Steelmaking: Theory and Practice, PHI Learning, 2008.↩︎

8. EAF electrical systems: Furnace transformers for small EAFs (1–5 tonnes) typically have ratings of 1–8 MVA, delivering secondary voltages of 100–300 V at currents of several thousand amperes. The transformer must have tap-changing capability to adjust power during different phases of the melting cycle. See: Bowman, B. and Krüger, K., Arc Furnace Physics, Verlag Stahleisen, 2009.↩︎

9. Small-scale EAFs in the 1–10 tonne range are widely used globally in steel foundries, specialty steelmakers, and developing-country mini-mills. They are standard equipment produced by furnace manufacturers including Inductotherm, ABP Induction, SMS group, and numerous Chinese manufacturers. Operating principles scale linearly from large to small, though specific energy consumption increases somewhat at smaller scale due to higher surface-to-volume ratio. See: Steel Founders’ Society of America (SFSA) technical publications.↩︎

10. Power requirements for small EAFs: A 3-tonne EAF with a specific energy consumption of approximately 450 kWh/tonne and a tap-to-tap time of approximately 90 minutes draws approximately 450 × 3 / 1.5 = 900 kW average, with peak power during bore-down approximately 2–3 times average. Total connected power (transformer rating) of 1.5–3 MW is appropriate. Based on standard EAF engineering calculations.↩︎

11. Small EAF refractory campaign life: Smaller furnaces typically achieve shorter campaigns in terms of heats because the refractory volume is smaller relative to the slag-line erosion rate, but this is offset by lower throughput. Campaign life of 100–300 heats for a well-operated small EAF with commercial refractories is based on industry practice for steel foundry EAFs. See: Schacht, C.A. (ed.), Refractories Handbook, CRC Press, 2004.↩︎

12. DC EAF technology: DC furnaces offer approximately 30–40% lower electrode consumption, reduced flicker, and more uniform melting compared to AC furnaces. The trade-off is a more complex power supply (rectification) and bottom electrode system. At small scale (1–5 tonnes), DC furnaces have been commercially successful. See: Jones, J.A.T., “DC Arc Furnaces — Past, Present, and Future,” Electric Furnace Conference Proceedings, AIST.↩︎

13. EAF transformer ratings at small scale: A 3-tonne EAF transformer of approximately 2–5 MVA rating is comparable to medium distribution transformers in NZ’s power system. NZ has transformer repair and rewinding capability at several facilities, and some capacity for new transformer construction. The key challenge for an EAF transformer is the duty cycle (frequent high-current surges, mechanical stress from short-circuit forces) which requires robust construction. See: Doc #69 for NZ transformer capability.↩︎

14. Manual EAF electrode regulation: Before automated electrode regulation systems (developed in the 1960s–70s), EAF operators manually controlled electrode position by observing ammeter readings and listening to arc sound characteristics — an experienced operator can judge arc stability by ear. Manual operation is less efficient and produces more electrode breakage, but it is feasible. See: historical EAF operating practice in Fruehan (note 2).↩︎

15. Induction melting fundamentals: The principle of induction heating — alternating magnetic fields inducing eddy currents that generate resistive heating — was first applied to metal melting in the early 20th century. The coreless induction furnace was developed commercially in the 1920s. See: Lupi, S., Fundamentals of Electroheat: Electrical Technologies for Process Heating, Springer, 2017; Patel, A.D. (ed.), “Induction Furnace — A Review,” International Journal of Engineering and Technical Research, 2015.↩︎

16. Induction furnace coil copper requirements: A 1-tonne coreless steel melting furnace typically has a coil made from 20–40 turns of 25–50 mm square copper tubing, with total coil weight of approximately 500–1,000 kg depending on design. Water flows through the tubing at approximately 5–10 litres per minute per turn to remove the I²R heat generated in the coil. Based on induction furnace design references: Simpson, P.G., Induction Heating: Handbook of Design and Applications, McGraw-Hill, 1960; manufacturer technical data from Inductotherm and ABP Induction.↩︎

17. Frequency and furnace size relationship: For effective electromagnetic coupling (penetration depth matching the charge dimensions), lower frequencies work with larger charge diameters and higher frequencies with smaller charges. At 50 Hz (mains frequency), effective coupling requires charge diameters above approximately 500–700 mm, corresponding to furnace capacities above roughly 3–5 tonnes for steel. Smaller furnaces need medium frequencies (500–3,000 Hz) for efficient operation. See: Lupi (note 14); induction heating engineering references.↩︎

18. Induction furnace crucible life: For coreless steel melting furnaces with rammed magnesia crucibles, typical campaign life is 80–200 heats, depending on steel chemistry, operating temperature, and crucible quality. With NZ-produced magnesia (from seawater — if developed), crucible life may be at the lower end of this range initially. Based on foundry practice data; Inductotherm crucible life guidelines.↩︎

19. Induction furnace coil design: Coil design follows standard electromagnetic induction calculations — the number of turns, coil geometry, operating frequency, and power input determine the magnetic field intensity and therefore the heating rate. Design procedures are well-documented in induction heating engineering texts. See: Simpson (note 15); Rudnev, V. et al., Handbook of Induction Heating, CRC Press, 2017.↩︎

20. Motor-generator set vs. solid-state inverter for induction furnace power supply: Efficiency figures based on typical induction furnace power supply performance data. Motor-generator sets for induction heating service (alternators driven by induction motors) achieve approximately 80–88% combined efficiency depending on load factor and machine design. Modern IGBT-based solid-state inverters achieve 93–97% efficiency. Frequency capability: purpose-built alternators for medium-frequency induction service typically output 500–1,000 Hz; standard industrial generators cannot readily exceed this. See: Rudnev et al. (note 18); Simpson (note 15); general induction heating power supply engineering references.↩︎

21. Soderberg electrodes versus graphite: Soderberg electrodes have approximately 5–10 times the electrical resistivity of graphite electrodes, resulting in higher electrode heating (I²R losses), more frequent electrode breakage, and higher consumption rates per tonne of steel. However, they avoid the need for the graphitisation step (2,500–3,000°C in an Acheson furnace), making them producible from NZ materials. Soderberg technology was standard in ferroalloy and aluminium smelting for decades. See: Soderberg electrode literature; Pratt, C.J., “Carbon and Graphite Products,” in Kirk-Othmer Encyclopedia of Chemical Technology.↩︎

22. Carbon electrode production: Amorphous carbon electrodes are produced by mixing calcined coke, anthracite, or charcoal with coal tar pitch, forming by extrusion or pressing, and baking at 800–1,200°C. They are less expensive but have much higher resistivity than graphitised electrodes. Used historically in some chemical process furnaces and still used in certain specialised applications. See: Marsh, H. and Rodríguez-Reinoso, F., Activated Carbon, Elsevier, 2006; general carbon electrode manufacturing literature.↩︎

23. Dolomite refractory hearth life: Rammed dolomite bottoms in small EAFs typically achieve 50–300 heats depending on steel chemistry, tapping method, and slag practice. Dolomite is a basic refractory that resists attack by basic slags but degrades through hydration if exposed to moisture. NZ has adequate dolomite deposits for this application. See: Schacht (note 10); NZ Crown Minerals data on dolomite deposits.↩︎

24. The Dow seawater magnesia process, commercialised in the 1940s in Freeport, Texas, precipitates magnesium hydroxide from seawater by adding lime (CaO) or dolime, filters and washes the precipitate, and calcines it to produce magnesia (MgO) suitable for refractory and other applications. Seawater contains approximately 1,300 mg/L of magnesium. See: Kramer, D.A., “Magnesium, Its Alloys and Compounds,” USGS Open-File Report 01-341; general industrial chemistry references.↩︎

25. Lime-based EAF slag practice without fluorspar: Fluorspar (CaF₂) is added to EAF slag to improve fluidity and promote sulfur removal. Without fluorspar, slag viscosity is higher and desulfurisation is less effective. For most structural steel applications, the resulting sulfur levels (0.030–0.050%) are acceptable, though higher than modern clean steel practice (0.010–0.020%). NZ has no known commercial fluorspar deposits. Based on EAF slag chemistry practice; Turkdogan, E.T., Fundamentals of Steelmaking, Institute of Materials, 1996.↩︎

26. NZ manganese deposits: Small manganese deposits are known in Northland (e.g., Waihopo, near Kaikohe) and the East Cape region, but none have been commercially mined at significant scale. Total identified resources are modest. Whether they could support even small-scale ferromanganese production is uncertain and would require geological assessment. See: GNS Science mineral occurrence databases; Crown Minerals records. https://www.nzpam.govt.nz/↩︎

27. Materials testing capability in NZ: HERA (Heavy Engineering Research Association) maintains testing facilities in Auckland. University of Auckland and University of Canterbury have metallurgical testing laboratories with tensile testing machines, Charpy impact testers, hardness testers, and optical emission spectrometers for chemical analysis. Several private testing firms (e.g., Metals NZ, Bureau Veritas) also provide testing services. This capability is essential for quality control of regionally produced steel. See: HERA. https://www.hera.org.nz/↩︎

28. NZ annual electricity generation: NZ generates approximately 42,000–44,000 GWh per year from predominantly renewable sources (hydro, geothermal, wind). See: MBIE, New Zealand Energy in Brief (annual publication); Electricity Authority, EMI — Electricity Market Information, generation statistics. https://www.mbie.govt.nz/energy/; https://www.emi.ea.govt.nz/↩︎

29. Voltage flicker from EAF operation: Rapid, irregular changes in arc current cause voltage fluctuations on the supply network (flicker). This is more pronounced for AC furnaces and at higher power levels. At the 1–3 MW scale, flicker is typically manageable on strong industrial supply points. Static VAR compensators (SVCs) can mitigate flicker but are complex power electronic systems. See: International Electrotechnical Commission (IEC) standards on flicker; Bowman and Krüger (note 7).↩︎

30. EAF cooling water requirements: Water-cooled panels, electrode holders, and (for induction furnaces) coils require recirculating cooling water. Total heat rejection for a 3-tonne EAF is approximately 0.5–1 MW thermal, requiring a cooling tower and approximately 10–20 m³/hour of makeup water. Based on standard EAF and induction furnace cooling system design.↩︎

31. Key reference texts for EAF and induction furnace operations: The Making, Shaping and Treating of Steel (AISE/AIST, multiple editions — the standard reference for all steelmaking processes); Fruehan, R.J. (ed.), Steelmaking and Refining Volume, AISE, 1998; Electric Furnace Steelmaking, I&S/AIST proceedings (annual conference publications); Simpson, P.G., Induction Heating: Handbook of Design and Applications, McGraw-Hill, 1960. NZ university libraries (University of Auckland, University of Canterbury) should hold copies of these standard texts. HERA’s technical library is another likely source.↩︎

32. Submerged-arc furnace technology: SAFs are the standard industrial process for producing ferroalloys (ferrosilicon, ferromanganese, ferrochrome), calcium carbide, and other electrometallurgical products. The electrodes are submerged in the charge (a mix of ore, carbon, and flux), and the arc heats the charge internally. Power levels range from <1 MW (pilot scale) to >100 MW (large industrial). See: Gasik, M. (ed.), Handbook of Ferroalloys, Butterworth-Heinemann, 2013.↩︎

33. Soderberg electrodes in SAFs: Soderberg electrodes were originally developed for submerged-arc furnace applications (Norwegian ferroalloy industry, early 20th century) and are still widely used in ferroalloy and carbide production globally. The buried-arc operation is more forgiving of electrode quality variations than open-arc EAF steelmaking, making SAFs a better match for NZ-produced Soderberg paste. See: Gasik (note 30); historical ferroalloy industry literature.↩︎

34. NZ acetylene production: Industrial acetylene in NZ is produced by BOC (now Linde) and Air Liquide by reacting calcium carbide with water at their gas filling stations. The calcium carbide is imported, primarily from China. NZ does not currently produce calcium carbide. When imported carbide stocks are exhausted, NZ’s oxy-acetylene capability ceases unless domestic carbide production is established. See: BOC/Linde NZ operations; Doc #3 for detailed analysis.↩︎

35. Calcium carbide SAF scale estimate: Commercial calcium carbide SAFs consume approximately 2,800–3,200 kWh of electricity per tonne of carbide produced. A 2–5 MW furnace operating at 85% availability produces approximately 2,100–4,500 tonnes per year. NZ’s pre-event acetylene consumption is not publicly reported in detail, but can be estimated from industrial gas supplier data and the volume of welding equipment in use; recovery-era demand will be substantially lower than peacetime. This estimate assumes carbide is the primary acetylene source and accounts for the lower overall industrial activity level in a recovery scenario. The figure requires verification against NZ industrial gas market data. See: Gasik (note 30); calcium carbide production data from CRU Group industrial minerals reports.↩︎

36. Parengarenga Harbour silica sand: The silica sand deposit at Parengarenga Harbour in the Far North of the North Island is one of the world’s highest-purity silica sand resources, with SiO₂ content exceeding 98%. It has been exported for glass and ceramics manufacturing. This resource is directly suitable for ferrosilicon and silicon metal production. See: Crown Minerals; GNS Science. Note: the Parengarenga area is culturally significant to Te Aupōuri iwi, and any industrial use of the sand resource would require consultation and partnership with local iwi.↩︎

37. Ferrosilicon production is one of the simplest SAF processes — the chemistry (SiO₂ + 2C → Si + 2CO, with iron as a collector) is straightforward, the raw materials (silica sand, charcoal, iron/steel scrap) are widely available, and the furnace operation is well-documented. Global ferrosilicon production is approximately 10–12 million tonnes per year, with China, Russia, and Norway as major producers. NZ’s requirements (a few hundred tonnes per year) are a tiny fraction of global production, but the process scales down to very small furnaces (pilot plants as small as 100 kVA have produced ferrosilicon successfully). See: Gasik (note 30).↩︎

38. Silicon metal production: Metallurgical-grade silicon (MGS, 98–99% Si) is produced in SAFs similar to ferrosilicon furnaces but with no iron in the charge and higher operating temperatures. Solar-grade silicon (99.9999%+) requires further purification by the Siemens process (trichlorosilane distillation and chemical vapour deposition) — an extremely demanding industrial chemistry project that is a Phase 5+ or later goal for NZ. See: Doc #115 (Semiconductor Roadmap) for the full silicon purification pathway; Schei, A., Tuset, J.K., and Tveit, H., Production of High Silicon Alloys, Tapir Forlag, 1998.↩︎