Hot Metal
Three things worth your attention this week.
Salzgitter has signed Tenova to build Germany's largest EAF at Hüttenwerke Krupp-Mannesmann: a 285-tonne-tap Consteel rated at 2.5 million tonnes a year, construction starting this August, completion set for 2029. The published pitch is scrap preheating and productivity. The quieter fact is that a Consteel is designed around a permanent liquid heel — the machine works because the furnace is never empty. This week's issue is about that iron.
Accounting departments periodically discover the hot heel and ask why tens of tonnes of iron are "held in inventory" at bath temperature. It's a fair question and it deserves a straight answer: that tonnage is process equipment made of iron — it buys feed-rate capacity, arc stability and nitrogen control. The wrong answer is no answer, because then the heel gets managed by whoever shouts loudest at month-end.
The European Commission published its iron and steel guidance for CBAM's definitive period on August 14: for steel products, only direct emissions count as embedded emissions for now. If your shop sells into Europe, the measurement discipline behind your emissions numbers just turned into commercial paperwork. Shops that treat those numbers like heat-log data will price them; the rest will absorb them.
The iron you don't tap is working
Every so often someone runs the numbers and notices that tapping the furnace closer to empty would "free up" tonnes of iron, and the hot heel goes on trial. In a DRI shop the heel wins every time. It's still worth rehearsing the defense, because a heel that's tolerated rather than managed hands the prosecution its evidence.
What the heel does, in the order the money is involved:
It melts your DRI. Continuous feeding only works because pellets dissolve into a superheated liquid bath; the bath is the melting machine. Feed onto a heel that's too small and pellets arrive faster than they dissolve — first the iceberg, then the fused mass that costs a shift of oxygen lancing to get out. Published flat-bath practice spans a wide band, from the mid-teens as a percentage of tap weight up to a large permanent heel in designs built around one. The band is wide because the right number is furnace-specific. The floor isn't negotiable.
It steadies the arc. A flat liquid bath under foamed slag is the calmest electrical environment an EAF can offer: stable arcs, damped flicker, power on continuously instead of the bore-in gymnastics of a bucket-charged heat. A good part of the flat-bath route's published kWh and electrode numbers is really the heel's performance, uncredited.
It keeps nitrogen honest. The heel sustains a continuous carbon boil, and CO bubbles are the only nitrogen exit a furnace has. Add low-nitrogen DRI displacing nitrogen-bearing scrap and you get the published split — scrap-route EAFs broadly in the 70-110 ppm band, high-DRI flat-bath shops reporting 30-60 ppm. For flat products sold against a tight nitrogen spec, the heel is quietly part of the product.
It's a thermal flywheel. Tens of tonnes of superheated iron soak up every disturbance — a feed hiccup, a delay, a cold ladle — that would otherwise show up at tap as a temperature miss.
Now the prosecution's honest points, because they're real. The heel carries history: slag and phosphorus from the last heat sit in the next heat's opening balance, and a shop chasing low-P grades has to run its slag-off discipline with that in mind. The heel occupies capacity — iron held in the furnace is tap weight you don't sell today, which is exactly why it has to earn its keep through the list above. And the sharpest point: the heel drifts. Nobody weighs it. It's inferred, charge weights in against tap weights out, and when tap weights wander the heel quietly grows or shrinks until it announces itself — as an over-full furnace in one direction, a feeding problem in the other. In the shops we've seen get bitten, the heel hadn't been reconciled in weeks. The drift was sitting in the data the whole time. Nobody owned it.
So the argument lands here: run the heel like controlled inventory. A target band with alarms on it, a charge-minus-tap reconciliation on a fixed cadence, and a named owner on every crew. A managed heel pays for itself every heat. An unmanaged one pays for itself until the week it doesn't.
Operator's Notebook — heel decision tree for a tap-weight deviation
Tap weight came in off target. Walk the tree before the next heat, not after three more.
Tap weight HIGH (heel shrinking)
Was it a deliberate over-tap — a ladle weight request, a sequence need?
Yes → log it; rebuild the heel over the next 2-3 heats by trimming tap weight. Never in one heat.
No → check EBT closure timing and slag detection at tap. A late close bleeds heel every heat and reads as "good yield" right up until feeding degrades.
Feeding symptoms already showing (slow dissolution, unmelted islands, arc instability at feed start)?
Yes → treat the heel as below floor: cut feed rate now, rebuild deliberately, investigate afterwards. Sequence pressure does not overrule the iceberg.
Tap weight LOW (heel growing)
Charge mass balance: did charged units (DRI + scrap + flux iron) match plan?
No → weighing or metallization input error. Fix the number before touching practice.
Yes → look for an early EBT close or slag-avoidance tapping habits. A cautious tapper grows the heel one careful heat at a time.
Furnace level visibly high, or door-threshold interactions?
Yes → schedule a controlled heel-trim tap. Don't absorb it by quietly under-charging; that hides the cause.
Either direction: update the running heel estimate at handover — charge-in minus tap-out, cumulative — and hand the next crew a number, not an impression.
Next week: carbon in DRI is not a contaminant. It's the cleanest fuel in the shop.
Written by active DRI-EAF operators. Anonymous by necessity, specific by design.