Hot Metal
Three things worth your attention this week.
Geopolitics just got into your electrode line. GrafTech filed an anti-dumping petition with Brazil's trade body DECOM in March; the agency opened an investigation with preliminary dumping margins of 54.9% on Chinese and 57.3% on Indian large-diameter (≥350 mm) electrodes. If you've been quietly leaning on cheap imported graphite, expect price to start splitting by origin. Model it before the next reline, not after — the cheapest tonne of graphite is still the one you never sublimate.
Stegra cleared a funding gate. A €1.4 billion round landed in April — finalizing through June — and management says construction now accelerates after a fundraising-focused stretch. Boden was reported around 60% complete last fall, but the production timeline is back under review, with delays acknowledged and key electrolysis kit still to install. The build's on the front foot; the date isn't pinned. What that first heat will and won't prove is its own story — we get into it next week.
Whyalla's restart is an opex bet, not a capex one. Australia has now put more than A$2.8 billion behind the works — the bulk of it earmarked for the green-iron transition — after they went into administration; M Resources and Jindal Steel were shortlisted in late May, with a Santos gas deal underpinning a DRI transition from 2030. Every brownfield EAF/DRI conversion we've watched teaches the same lesson: announced capex is the easy part. Feed, power, and electrode opex decide whether the plant sees year three.
Why kg/t hides the real electrode story
Almost every shop benchmarks graphite electrodes on one number: kg per tonne of liquid steel. It's the number on the monthly cost sheet, the number the GM compares against last quarter, and the number vendors quote when they sell you a reduction. It's also the number that hides where your money is actually leaking.
A single kg/t figure — for a typical EAF, somewhere in the 2-3 kg/t band the market reports still quote — blends at least three physically different mechanisms. Each has a different driver and a different fix. Treat them as one line and you're optimizing blind.
Tip consumption is the electrode sublimating and oxidizing at the arc. It scales with the electrical work you do: current, arc energy, power-on time. It's the part the electrode salesman talks about, because it's the part a better grade of graphite genuinely improves. But in a well-run UHP furnace — and UHP is more than half of all electrode demand now — it's rarely the biggest slice.
Side oxidation is the column burning off above the bath. Chemical attack from the furnace atmosphere, radiation on the exposed graphite. It's driven by exposed length, time at temperature, and how much air gets in. The published consumption literature puts oxidation around two-thirds of total electrode loss, with the cylindrical side surface alone around half — more under heavy oxygen lancing. Graphite grade has almost nothing to do with it. Furnace sealing, delay time, and how long the column sits hot between heats have almost everything.
Breakage and spalling is the mechanical loss. Scrap cave-ins on a charge furnace, thermal shock, joint failures, a regulation overshoot driving a tip into a cold pile. Episodic, ugly, almost entirely operational. On continuous DRI feed the cave-in mode mostly disappears — but long, uninterrupted power-on quietly loads the tip and side instead. Different furnace, same trap.
This is also why the "20-40% electrode reduction" claims now attached to water-cooled composite electrodes and similar kit deserve one question before the PO: reduction from what baseline? A 30% cut off a shop running side oxidation hot is a different purchase than 30% off an already-tight operation. Only the second one is worth what it costs.
And here's the part that doesn't make the vendor deck. Crew-to-crew variance on the same furnace, same electrodes, same feed routinely dwarfs the gain from a capital optimization project. Two crews running the identical shop show meaningfully different kg/t — regulation discipline, how aggressively they bore in, foamy-slag management, how long they leave the column hot during a delay. We've seen this hold across every shop we've worked in. If the spread between your best and worst crew is two or three times what a reduction project buys, the cheapest electrode reduction you'll ever get is a shift-handover standard, not a purchase order.
So what do you measure instead of one number?
Split the bill. Estimate tip from an electrical/length model, side oxidation from exposed-length × power-on time, and log breakage as discrete events. You don't need lab precision — you need to know which slice is growing.
Benchmark by crew, not just by month. The monthly average launders the variance you most want to see.
Track per-column. A single phase running consistently shorter is telling you about regulation balance or hidden side oxidation long before the monthly kg/t moves.
A single kg/t number isn't wrong. It's an average of three problems, and you can only fix them one at a time. The shops that get electrode cost down and keep it down stopped managing the average years ago.
Operator's Notebook — 5 heat-level signals most dashboards skip
Things worth a glance at handover that the standard panel won't flag:
Acoustic–electrical foam agreement. When the acoustic foam index and the electrical stability reading disagree, trust neither — the slag isn't doing what the panel thinks it is. Disagreement usually means foam is collapsing somewhere the sensor isn't looking.
Feed-point drift on continuous DRI feed. Watch where the feed is actually landing relative to the hot heel over a shift. A wandering feed point changes melt-in and carbon pickup before any setpoint shows it.
Per-phase electrode length differential. One phase consistently shorter than the other two is a regulation-balance or uneven-wear signal. Catch it here, not on the month-end consumption report.
Power-on vs. power-off ratio per heat. Delay time isn't free — every minute the column sits hot off-power feeds side oxidation and refractory exposure. That cost never shows on the kg/t line, but it's in the bill.
Foam height stability, not just presence. Foam that's there but swinging spikes radiation onto the panels and the column on every collapse. Variance is the enemy, not absence.
None of these need new capital. They need someone on the floor watching the right thing and writing it on the board.
Next week: what Stegra's first heat will actually prove — and the H2-DRI cost number that decides everything.
Written by active DRI-EAF operators. Anonymous by necessity, specific by design.