Technical Article
A panel's material is decided by its position in the furnace, not by its type. A comparison calculated from published data.
Published: October 21, 2026
Short answer: It is not the panel type that decides, but its position in the furnace.
Under normal operation a wall panel sees a heat flux of 600–700 kW/m². At that load the steel wall reaches only 126 °C — comfortably safe for a material rated to 400 °C. This is why most of the furnace wall uses steel panels.
Panels facing the oxygen injectors are different. When a jet blows back, they see 4.000–6.000 kW/m². At 6.000 kW/m² the temperature difference across the steel wall alone is 462 °C; the 400 °C limit is exceeded before the water side is even considered. Under the same conditions the difference across a copper wall stays at 71 °C.
So the question is not "which material is better" but "where will this panel be mounted".
Below we show where those figures come from.
Panel selection starts with knowing the load. Toulouevski and Zinurov compiled measured maximum heat flux densities for water-cooled furnace elements. The three values relevant to wall panels are:
| Wall panel condition | qmax (kW/m²) |
|---|---|
| Heavily encrusted with slag | 200 – 300 |
| Normal operation, depending on arc shielding | 600 – 700 |
| At oxygen jet blow-back | 4.000 – 6.000 |
Source: Toulouevski & Zinurov, Innovation in Electric Arc Furnaces, 2nd ed., Springer, Section 12.3.1.
The third row needs a brief explanation.
In modern electric arc furnaces, oxygen is injected into the bath through injectors mounted in the sidewall to speed up melting and refining. When the jet strikes the scrap or the liquid bath, part of it is deflected back toward the wall — the industry calls this blow-back. At that moment the panel near the injector takes a short but very intense heat pulse.
The authors treat this as a separate operating condition: the same wall panel sees 4.000–6.000 kW/m² instead of 600–700 kW/m² during blow-back. The panel does not change; the condition does. Panels opposite burner and injector ports are exposed to this event; those further away are not.
The authors' warning governs the whole design approach: elements must be sized for the peak flux, not the time-averaged one. Peak fluxes are brief but recur; they determine element durability and cause premature failures. A sufficiently high peak flux can strip the skull from the surface within minutes, melt the surface, or cause burn-through.
The material decision therefore follows the peak load of the mounting position, not the average load of the panel.
Toulouevski and Zinurov give a complete worked calculation for a water-cooled pipe (Example 3, Chapter 3):
Water flows at 3,0 m/s in a low-carbon steel pipe with d₂ = 60 mm, d₁ = 52 mm, wall thickness δ = 4 mm. Water temperature is 30 °C. The heat flux density to the water is q = 600 kW/m².
Their calculation: the water-side convection coefficient α = 11,5 kW/m²·°C, giving an inner wall temperature of
t1 = 30 + 600/11,5 = 82 °C
The outer (hot face) wall follows from t2 = t1+ q·δ/k, with the book's generic low-carbon steel value k = 55 W/m·K:
t2 = 82 + 600.000 × 0,004 / 55 = 126 °C
Repeating the same calculation and changing only the wall material:
| Material | Thermal conductivity | Hot face temperature |
|---|---|---|
| Low-carbon steel | 55 W/m·K | 126 °C |
| Cu-DHP | 339 W/m·K | 89 °C |
| Cu-ETP | 394 W/m·K | 88 °C |
(Conditions: q = 600 kW/m², water at 3,0 m/s and 30 °C, 4 mm wall. Water-side conditions held constant; only the wall material was changed.)
The result is initially surprising: at 600 kW/m² the steel wall reaches only 126 °C.
P265GH boiler tube steel has a continuous service range of −10 … +400 °C, and the proof strength table in the material standard ends at 400 °C (Nirotec material data sheet, 1.0425). The limit is not a sharp failure point but a progressive weakening: proof strength (Rp0,2) falls from 220 MPa at 100 °C to 153 MPa at 400 °C — roughly a 30 % loss. A panel is a pressure vessel operating under water pressure, and that loss comes straight out of the safety margin.
126 °C is far below that band. This is why most wall panels are steel.
The real question is not the normal load but what happens at the moment of blow-back.
Here the water-side calculation becomes uncertain — at around 4.000 kW/m² the water enters the boiling regime and the simple convection formula no longer holds. But the temperature drop across the wall is independent of that uncertainty; it is pure conduction, given by ΔT = q·δ/k:
| Heat flux | P265GH wall (4 mm) | Cu-DHP wall (4 mm) |
|---|---|---|
| 600 kW/m² (normal) | 46 °C | 7 °C |
| 4.000 kW/m² (blow-back) | 308 °C | 47 °C |
| 6.000 kW/m² (blow-back) | 462 °C | 71 °C |
At 6.000 kW/m² the difference between the two faces of the steel wall is 462 °C on its own — the 400 °C service limit is exceeded without adding the water-side temperature at all. Under the same conditions the drop across copper stays at 71 °C.
This is where copper earns its cost. At normal load there is no practical difference between the two materials; during blow-back steel reaches its limit and copper does not. And as the book states, it is those brief moments that determine element life.
This is why a single furnace carries both steel and copper panels: copper at positions near burner and oxygen injector ports, steel everywhere else.
Skull and slag build up on the panel surface. This is not fouling — it is part of the design.
Toulouevski and Zinurov report that on wall panels heavily encrusted with slag the peak heat flux falls to 200–300 kW/m² — less than half the 600–700 kW/m² of an uncoated panel. The uncoated values correspond to the black-body radiation of heavily dust-laden gases at 1.700 °C, which is 680 kW/m².
The practical consequence: as long as slag adheres, the panel operates far below its design load.The material difference shrinks in this regime. Copper's advantage appears when the slag fails to adhere or is stripped away — and those are precisely the moments when elements fail.
This is why studs, cups or pins are welded to the panel face to promote slag adhesion.
| Cu-ETP | Cu-DHP | P265GH | |
|---|---|---|---|
| Thermal conductivity (20 °C) | 394 W/m·K | 339 W/m·K | 52 W/m·K |
| Electrical conductivity | 58 MS/m (100 % IACS) | 50 MS/m (85 % IACS) | — |
| Tensile strength | — | — | 470 MPa |
| Proof strength | — | — | 260 MPa |
| Density | 8,92 g/cm³ | 8,9 g/cm³ | 7,85 g/cm³ |
| Specific heat (20 °C) | 394 J/kg·K | — | 470 J/kg·K |
| Melting point | — | 1.083 °C | — |
| Continuous service temperature | — | — | −10 … +400 °C |
| Proof strength at 400 °C | — | — | 153 MPa — a 30 % loss from 220 MPa at 100 °C |
Sources: KME Cu-ETP data sheet · Aurubis C12200 Cu-DHP data sheet · Nirotec P265GH (1.0425) data sheet
In ratio terms, Cu-ETP conducts 7,6 times and Cu-DHP 6,5 times better than the steel.
The calculation above describes the steady state. But an EAF load is not steady — it varies through charging, melting, refining and tapping, and the arc can be deflected toward the wall when it burns unstably.
How quickly a panel responds to that change is governed by thermal diffusivity:
α = k ÷ (ρ · cp)
| Calculation | Thermal diffusivity | |
|---|---|---|
| Cu-ETP | 394 ÷ (8.920 × 394) | 112 mm²/s |
| P265GH | 52 ÷ (7.850 × 470) | 14,1 mm²/s |
Copper responds to a temperature change 8 times fasterthan steel. This is copper's second advantage in zones dominated by cyclic thermal shock: when a peak flux arrives, the surface temperature rises less, the cycle amplitude narrows, and fatigue cracking starts later.
(Density and specific heat values are taken from the sources in the table above.)
This distinction is critical in panel manufacture, yet it is rarely written about.
Cu-ETP (C11000) is the best-conducting copper at 394 W/m·K. However, it is susceptible to hydrogen embrittlement in reducing atmospheres, which becomes a problem during welding.
Cu-DHP(C12200) is phosphorus-deoxidised copper. Its 0,015–0,040 % phosphorus binds the oxygen and eliminates hydrogen embrittlement; the Aurubis data sheet rates its weldability as "excellent". The price is conductivity dropping from 394 to 339 W/m·K — roughly a 14 % loss.
The consequence ties the manufacturing method to the material choice:
| Manufacturing method | Copper grade | Reason |
|---|---|---|
| Welded panel | Cu-DHP | No hydrogen embrittlement risk at the weld |
| Cast / non-welded | Cu-ETP | No constraint; full conductivity available |
DNL uses Cu-DHP in its welded panels and Cu-ETP in other work; the choice is made per project.
So "copper panel" is not a sufficient specification. When requesting a quotation, ask which grade is used and whether the panel is welded or cast — the two are linked.
The values below are drawn from a 1989 doctoral study and reflect the panel practice of that period. Panel technology has advanced since; read these as orders of magnitude.
| Panel type | Life (number of heats) |
|---|---|
| Steel box panel (Korf/Fuchs type) | at least 1.500, typically 1.000 more |
| Cast cooling block | 250 – 5.000 |
| Copper panel | over 10.000 |
| Copper panel — lower wall hot spot | 3.300 |
Source: Simon, M.J., The thermal performance of water cooled panels in electric arc steelmaking furnaces, PhD, 1989, p. 11.
The last row matters: a copper panel's life falls from 10.000 to 3.300 heats when placed at a hot spot. Copper wears out in the hardest position too — only more slowly than steel.
Most panels fail by cracking, with hot-face shrinkage and cyclic thermal shock given as the causes. Other failure modes are arcing onto the panel, careless oxygen lancing, partial immersion in liquid steel, and scrap impact (ibid., p. 10).
Copper is not used everywhere because of cost. It is significantly more expensive than steel in both raw material and labour, so its use is confined to positions where the gain in service life justifies the expense.
Whatever the panel material, it will fail if the water side is wrong.
Toulouevski and Zinurov give the minimum water flow rate relation:
V = Q ÷ [1,15 × (toutlet − tinlet)] , m³/h
where Q is the total heat flux received by the element during its most thermally stressed period (kW).
The temperature limits are critical:
| Condition | Maximum outlet water temperature | Reason |
|---|---|---|
| Chemically untreated, low-quality water | 40 – 50 °C | Salt deposits on the channel walls |
| Chemically treated water | 70 °C | Above this, inner and outer wall temperatures also rise |
Salt deposits reduce heat transfer, which raises the wall temperature, which accelerates deposition. It is a self-reinforcing loop and a common but late-detected cause of panel failure.
On water velocity, supplier practice clusters in the 1,2–2,5 m/s range; velocities above 5 m/s have been reported for high-load zones such as the lower slag line (ispatguru). The book's worked example uses 3,0 m/s.
Water cooling means taking heat out of the furnace — and that heat is energy.
Toulouevski and Zinurov give a detailed heat balance for a modern 130-tonne EAF. The furnace operates with a 38-minute tap-to-tap time, 78 MW power and 331 kWh/tonne electricity consumption.
In that balance, the heat going to the water-cooled wall and roof panels is 77 kWh/tonne, or 11,4 % of total energy input (Section 4.3). For comparison, the enthalpy of the liquid steel — the useful part of the work — is 58,1 %.
So in a modern furnace roughly one eighth of the energy input leaves through the water-cooled panels.
This looks like a loss item, but there is another side to the trade: a water-cooled wall substantially reduces refractory consumption and downtime. Operating at the same power without panels is not possible.
The data above comes from industry literature. What follows is DNL's own manufacturing data.
| Water-cooled steel panels | P265GH-TC1 (seamless/drawn), P235GH-TC1, A106 Gr.B and equivalents |
| Water-cooled copper panels | 99 % copper — Cu-DHP in welded manufacture, Cu-ETP in other work |
DNL manufactures panels in both materials; a single furnace can be supplied with a mix determined by position. Related components: EAF upper shell, EAF roof, water-cooled elbow.
| Diameter range | Wall thickness | |
|---|---|---|
| Copper | Ø76 – Ø90 mm | 10 – 12 mm |
| Steel | Ø33 – Ø114 mm | 4 – 15 mm |
The book's worked example uses a Ø60 mm steel pipe with a 4 mm wall — within this range.
| Furnace capacity | up to 100 tonnes (delivered units in the 50–100 tonne range) |
| Upper shell diameter range | Ø3.000 – Ø10.000 mm |
| Hydrostatic test pressure | 12 – 40 bar to customer requirement (typically 12 – 20 bar requested) |
| Welding standard | EN ISO 3834-2 |
| Non-destructive testing | Standard: penetrant (PT) + radiographic (RT). On request: ultrasonic (UT), magnetic particle (MT) |
| Warranty period | 12 months |
A common misconception about pressure is worth correcting: a copper panel does not require a separate pressure class. Copper and steel panels in the same furnace are connected to the same cooling circuit and operate at the same pressure. Carrying both materials in one furnace is normal practice. The 12–40 bar above is a hydrostatic test pressure, applied to verify manufacturing integrity; it should not be confused with the operating pressure.
DNL manufactures both steel and copper panels. This makes the material choice a technical rather than a commercial decision, because there is no single product to sell.
