A single EAF refractory campaign runs anywhere from half a million to two million dollars depending on furnace size and lining specification, and the decision to end that campaign early or push it a few heats too far is still made mostly on visual inspection and operator gut feel. Corner thinning, hot spots near the slag line, and localized slag attack rarely show up as a uniform wear pattern, which means a lining that looks fine in most zones can be dangerously thin in one. Our refractory engineering team can review your last few campaign logs against what a wear prediction model would have flagged and when.
Steel Making — EAF Refractory
Stop Guessing Where the Lining Is Thin
Slag attack, thermal cycling, and mechanical wear don't erode a furnace lining evenly. An AI model tracks wear zone by zone across every heat, so campaign decisions are based on the thinnest point, not an average guess.
Wear Zone Snapshot
Hot Spot
Corner 3
Slag Line
Stable
Sidewall
Stable
Bottom
Low Wear
Slag Line
Stable
Hot Spot
Corner 1
Two of six zones flagged for accelerated wear this campaign
The Problem With Uniform Wear Assumptions
Most refractory replacement decisions still rely on a single campaign-life number derived from average brick consumption across the whole furnace, applied as a blanket target heat count before relining. In practice, wear is never uniform: corners near the electrode arcs see disproportionate thermal cycling, the slag line takes the brunt of chemical attack, and specific hot spots emerge based on scrap charging patterns and arc flare exposure that shift from campaign to campaign.
Treating campaign life as a single number means the furnace is often relined either too early, sacrificing usable life across zones that were still in good condition, or too late, running with a dangerously thin section in one corner while the rest of the lining looks acceptable on visual inspection.
$500K–$2M
typical cost per full refractory campaign, by furnace size
15-20%
of campaign life often left unused from early blanket relining
6
typical distinct wear zones tracked around an EAF shell
What Drives Localized Wear, Zone by Zone
Each zone of the furnace lining wears according to a different combination of mechanical and chemical stress, which is why a single campaign-average model misses the zones that actually determine when relining becomes necessary.
Corner Thinning
Electrode arc flare and thermal cycling concentrate stress at the hot-spot corners nearest the arcs.
Slag Line Attack
Chemical attack from foamy slag chemistry erodes the band at the slag interface fastest.
Sidewall Erosion
Scrap impact during charging causes mechanical wear that varies by charging pattern and bucket sequencing.
Bottom Wear
Generally the slowest-wearing zone, but tap hole erosion can accelerate wear locally over a campaign.
Want to see which zones on your furnace are wearing fastest right now?
Book a walkthrough and we'll map your last campaign's wear data.
Building a Zone-by-Zone Wear Model
Rather than tracking a single brick consumption average, a wear prediction model ingests heat-by-heat data that correlates with localized stress, including scrap charge patterns, arc power distribution, slag chemistry, and tap-to-tap cycle time, then maps predicted wear rate against each of the furnace's distinct zones separately.
Early campaign: baseline wear rate established per zone from first weeks of heats
Mid campaign: zones diverge from baseline as chemistry and charging patterns accumulate stress
Late campaign: thinnest zones flagged individually against a safety threshold, not a shell average
Relining decision: made from actual remaining thickness by zone, with lead time to plan the outage
Gunning and Maintenance Get More Targeted Too
Zone-level wear tracking doesn't only inform the final relining decision, it also changes how gunning and hot repair are scheduled during the campaign. Instead of a blanket gunning pass applied on a fixed interval regardless of actual condition, maintenance crews can target the specific zones the model flags as wearing fastest, applying repair material where it actually extends usable campaign life rather than spreading a fixed material budget evenly across zones that didn't need it.
| Approach | Relining Trigger | Gunning Strategy |
| Blanket campaign average |
Fixed heat count or scheduled outage |
Uniform gunning pass on fixed interval |
| Zone-level wear model |
Thinnest zone crosses safety threshold |
Targeted gunning on flagged zones only |
Why Campaign Extension Matters Beyond Material Cost
Every extra week a campaign runs safely past its previously assumed blanket life is a week without a relining outage, which on a shop running near capacity is often worth more in reclaimed production time than the refractory material savings themselves. Conversely, catching an accelerating hot spot early enough to schedule a planned partial repair, rather than an emergency shutdown mid-heat, avoids the far larger cost of an uncontrolled breakout event.
Extended Life
Full usable campaign life recovered from zones that weren't the limiting factor.
Planned Outages
Relining scheduled around actual condition instead of a fixed calendar interval.
Breakout Risk Down
Early warning on accelerating hot spots before they become safety events.
Frequently Asked Questions
What data is needed to build a zone-level wear model for our furnace?
A useful starting model typically draws on heat-by-heat records of scrap charging sequence, arc power distribution, slag chemistry, and tap-to-tap cycle time, ideally cross-referenced against periodic laser scanning or refractory thickness measurements taken during past campaigns. Furnaces that already collect this data through existing Level 2 systems can generally get a model running faster, while those with less historical measurement data start with a shorter lookback period and improve accuracy as more campaigns are logged.
Reach out to our team to review what your current historian already captures.
Do we need to install new sensors to get zone-level wear tracking?
Many shops already have enough process data in their existing historian to build an initial wear model without new hardware, since the correlating signals like arc power, charging pattern, and slag chemistry are typically already logged for other purposes. Adding periodic laser scanning or a permanent thickness monitoring system improves model accuracy meaningfully over process-data-only predictions, but it's an enhancement rather than a prerequisite for getting started.
Book a demo to see what accuracy level is achievable with your current instrumentation.
How does the model handle a change in refractory brand or specification?
A change in refractory specification is treated as a new baseline for the affected zones, since wear rate depends heavily on the specific material's resistance to thermal cycling and chemical attack, and the model needs a period of heats on the new material before its zone-level predictions reach full confidence. During this transition period the model typically relies more heavily on general wear pattern knowledge from similar specifications until enough campaign-specific data accumulates.
Talk to our team about how a specification change would be handled in your rollout.
Can this help decide when to do a partial hot repair versus a full reline?
Zone-level thickness prediction is specifically useful for this decision, since it separates zones that are approaching a safety threshold from zones still well within usable life, which is the information needed to judge whether a targeted hot repair on one or two zones can safely extend the campaign versus requiring a full shell reline. This decision still involves engineering judgment about repair feasibility and remaining structural integrity, but the model gives that judgment a data-backed starting point rather than a purely visual estimate.
Book a walkthrough to see a sample repair-versus-reline analysis.
How far in advance does the model give warning before a zone becomes critical?
Warning lead time depends on how quickly a specific zone is wearing, but the goal of continuous zone tracking is to surface an accelerating trend days to weeks before the zone would cross a safety threshold, rather than only flagging it once thickness measurements confirm a problem. This lead time is what allows maintenance teams to plan a targeted repair or schedule relining during a planned outage window instead of reacting to an unplanned shutdown.
Reach out to discuss typical lead times for furnaces similar to yours.
Extend Campaign Life With Confidence
Track Wear Zone by Zone, Not Shell by Average
Share your last two campaign logs and we'll show you where a zone-level model would have flagged accelerated wear before it showed up on inspection.
