A BOF refractory campaign is one of the most capital-intensive cycles in steelmaking, with each relining event pulling a converter offline for days and consuming millions in material, labor, and lost production. Despite that cost, most mills still assess lining condition through periodic manual inspections — opening the vessel during a scheduled stop, measuring a handful of points with a laser or thickness gauge, and extrapolating the overall wear profile from a fraction of the total surface area. The problem is that wear between those inspections is invisible, meaning a localized thin spot in the trunnion region or behind the charge pad can advance to a critical thickness without anyone knowing until the next scheduled check — or worse, until a breakout warning triggers an emergency stop. AI-powered camera and laser scanning systems change this by capturing lining imagery during normal vessel operations and building a per-heat wear map that makes deterioration visible as it happens, and a demo can show how that data translates into specific gunning and relining decisions.
BOF Refractory AI
BOF Refractory Life Monitoring with Camera Systems: Per-Heat Visibility Into Lining Wear
AI cameras and laser scanning capture BOF lining condition every heat — mapping wear zones, optimizing gunning, and extending campaign life with data instead of estimates.
What Drives BOF Refractory Consumption
Refractory cost in a BOF is not a single line item — it is the cumulative total of the initial lining installation, periodic gunning repairs, emergency patching, and the full relining event that ends each campaign. The largest single cost driver is usually the relining itself, both because of the direct material and labor expense and because the vessel is offline for the duration, which at high production rates represents substantial lost output. What makes this cost particularly frustrating for refractory leads is that a significant portion of it is driven by uncertainty: when you cannot see the lining between scheduled inspections, the default response is to gun more frequently and relining earlier than necessary as a safety margin against a breakout.
8-12%
of total BOF operating cost is refractory consumption including lining, gunning material, and relining labor
3-7 Days
average offline duration for a full BOF relining depending on vessel size and crew availability
15-30%
gunning material reduction reported by mills that shifted from scheduled to condition-based application
Wear Zones That Determine Campaign End
A BOF vessel does not wear evenly. Specific zones experience dramatically different combinations of chemical erosion, mechanical impact, thermal cycling, and slag exposure, which means campaign life is almost always determined by the single zone that reaches minimum thickness first — even if the rest of the lining has considerable remaining life. Understanding which zones degrade fastest and why is the foundation for any effective monitoring strategy, because a system that cannot see the critical zones is not providing the information that actually drives relining decisions.
Trunnion Region
Highest Combined Wear Rate
The area around the trunnion rings sees concentrated slag line contact, mechanical stress during every tilt cycle, and limited gunning access, making it the zone that most often forces campaign end in converters running high-phosphorus or high-manganese heats.
Charge Pad
Impact and Abrasion Zone
Heavy scrap strikes the charge pad during every heat, causing concentrated mechanical damage that varies with scrap practice, charge weight, and drop height. This zone wears in a pattern that accelerates unpredictably when scrap quality or charging procedure changes.
Knuckle Area
Thermal Cycling Transition
The transition between the cylindrical barrel and the conical bottom experiences high slag exposure combined with repeated thermal cycling, creating a wear pattern that accelerates in later campaign heats as the lining thins and thermal shock resistance drops.
Tap Hole Surround
Erosion From Steel Flow
High-velocity steel flow during tapping erodes the refractory around the tap hole, requiring frequent patching and periodic redrilling. Wear rate here depends heavily on tap time, stream angle, and lance positioning during the blow.
The Technology Stack Behind Per-Heat Monitoring
An AI-powered refractory monitoring system for a BOF is not a single camera pointed at the vessel interior. It is an integrated combination of imaging hardware, laser measurement, environmental protection, and a deep learning inference pipeline that converts raw sensor data into a structured wear map updated after every heat. Each component in the stack has a specific role, and a weakness in any one of them — poor camera positioning, inadequate housing protection, or an undertrained model — degrades the entire system's reliability.
1
High-resolution cameras mounted on the BOF hood or tilting mechanism capture interior imagery during charging, tapping, and slagging operations when the lining surface is exposed and visible.
2
Laser profilometers supplement visual data by measuring the distance to the lining surface at hundreds of points per scan, building a dimensional profile that quantifies remaining brick thickness across accessible zones.
3
A deep learning model processes each heat's imagery against a baseline established from the new lining, detecting changes in surface condition including crack propagation, spalling, exposed brick structure, and slag adhesion patterns.
4
Wear measurements are mapped to a zone-based vessel model, updating the predicted remaining life for each zone after every heat and generating alerts when any zone approaches its minimum thickness threshold.
Traditional Inspection vs AI-Powered Per-Heat Monitoring
The gap between these two approaches is not incremental — it is structural. Traditional inspection provides periodic snapshots that require extrapolation between data points, while per-heat AI monitoring generates a continuous data stream that makes the wear trajectory visible in real time. For a refractory lead making gunning and relining decisions, the difference between these two information sets is the difference between estimating and knowing.
| Assessment Aspect | Conventional Practice | AI-Powered Method | Measured Impact |
| Inspection Frequency |
Every 50-200 heats during scheduled stops |
Every heat during normal operations |
10-50x more data points per campaign |
| Surface Coverage |
20-40 discrete measurement points |
Continuous imaging of accessible lining |
Full-zone visibility versus spot sampling |
| Wear Trend Visibility |
Comparison between periodic snapshots |
Per-heat progression curve per zone |
Early acceleration detection versus after-the-fact discovery |
| Gunning Timing |
Fixed schedule or operator judgment |
Triggered by zone-specific thickness threshold |
Reduced over-gunning of healthy areas |
| Campaign End Decision |
Based on minimum measurement plus safety margin |
Based on probabilistic remaining-life model |
Potential 500-2000 additional heats per campaign |
Per-Heat Visibility
See BOF Lining Wear Update After Every Heat
A live demo shows how camera and laser data builds a real-time wear map your refractory team actually uses for gunning and campaign decisions.
From Wear Maps to Gunning Decisions
The operational value of per-heat wear data is not in the visualization itself — it is in the decisions that visualization enables. A wear map that stays on a screen without connecting to the actual gunning workflow is just a more expensive version of the laser scan report that already exists. The systems that deliver measurable ROI are the ones that close the loop between detection and action, turning raw thickness data into specific gunning instructions that crews can execute and verify.
1
The wear map identifies which zones have thinned below the gunning trigger threshold and which zones retain sufficient thickness, replacing blanket gunning schedules with zone-specific targeting that applies material only where needed.
2
Gunning crews receive a marked-up vessel diagram showing exact repair locations and target thickness, reducing execution variation between crews, shifts, and individual operators.
3
Post-gunning scans verify that repair material was applied to the correct zones at sufficient depth, creating a closed verification loop between the repair action and its measured result.
4
Gunning material consumption is tracked per zone per heat, giving the refractory lead a clear cost-to-remaining-life ratio that supports budget forecasting and supplier performance discussions.
Where Refractory Monitoring Deployments Fail
Not every AI refractory monitoring project delivers on its projected value, and the root causes are usually not failures of the AI technology itself but of the deployment decisions that surround it. The gaps below are the most common reasons an installed system underperforms, and each one is preventable with proper scoping before hardware is mounted and models are trained.
A
No Line-of-Sight to Critical Zones
Cameras mounted for mechanical convenience rather than optical coverage often cannot see the trunnion region or charge pad — the exact zones that determine campaign end — leaving the most important areas unmonitored.
B
No Integration With Gunning Scheduling
Wear data that remains in a standalone dashboard without connecting to the maintenance planning system means gunning decisions still rely on manual report review instead of automated threshold triggers reaching the right crew.
C
Model Trained on One Campaign Only
A wear model built from data collected during a single campaign may not generalize when steel grade mix, slag practice, or operating temperature changes between campaigns, causing accuracy to degrade without retraining.
D
Deploying Mid-Campaign Without Baseline
Installing the system on a vessel that is already mid-campaign means the AI has no reference imagery of the new lining, making early wear measurements less reliable until a full campaign cycle provides a proper baseline.
Frequently Asked Questions
How does AI camera monitoring differ from the laser scanning systems some mills already use?
Traditional laser scanning measures lining thickness at discrete points during scheduled vessel openings, generating a point cloud that represents a dimensional snapshot in time. AI camera systems complement this by capturing visual imagery during every heat, allowing the model to track surface changes — cracking, spalling, exposed brick structure, and slag adhesion — that pure distance measurement cannot detect. The combination of visual and dimensional data gives a more complete picture because a laser might show acceptable thickness while the camera reveals structural degradation that will accelerate wear in the next few heats.
A demo can show how visual and laser data combine in a single wear map.
Can AI refractory monitoring actually extend a BOF campaign, or does it just provide better documentation?
Extension comes from two mechanisms: first, by identifying and gunning localized thin spots before they force a campaign end, the system prevents a single high-wear zone from terminating a campaign while the rest of the lining still has significant remaining life. Second, by reducing unnecessary gunning on zones that are not approaching minimum thickness, the system preserves those areas for their full potential life rather than adding gunning layers that can spall or create uneven surfaces. Mills that have implemented condition-based gunning with per-heat data consistently report campaign extensions of 500 to 2,000 heats, which at typical BOF production rates represents substantial value in avoided relining costs and lost production.
Talk to a specialist about what campaign extension looks like for your vessel.
What environmental conditions affect AI camera performance inside a BOF?
The BOF environment is challenging for any optical system — high temperatures exceeding 1600 degrees Celsius during the blow, fume and dust from slagging, and intense radiant light all affect image quality. The most critical factor is having a clear enough view during the specific operations when the lining is visible: scrap charging, tapping, and slag pouring. Cameras need protection through water-cooled housings or air-purge enclosures, and the AI model must be trained to handle variable lighting and partial obscuration from fume. Systems that require perfectly clean images to function will fail in production, which is why model robustness to real-world imaging conditions matters more than performance on clean reference images alone.
Book a demo to see how the system handles real BOF imaging conditions.
How does per-heat monitoring change the refractory lead's daily workflow?
Instead of reviewing a laser scan report every few weeks and deciding whether to adjust the gunning schedule, the refractory lead has access to a continuously updated wear map showing exactly where the lining stands after the most recent heat. This shifts the role from reactive assessment — determining how bad the lining is right now — to proactive management — identifying which zones will need attention in the next 50 heats and ensuring material and crew are ready. The daily check becomes a quick review of exception alerts and trend changes rather than a full manual assessment, freeing time for supplier quality discussions, gunning procedure optimization, and campaign planning that were previously crowded out by data collection.
Support can walk through how the daily workflow changes with per-heat data.
What is the typical ROI timeline for a BOF refractory monitoring system?
Most mills see a return within one to two campaigns, driven by three value streams: gunning material reduction from eliminating over-gunning of healthy zones, campaign extension from catching localized wear before it forces relining, and avoided emergency stops from breakout prevention. The fastest payback comes in mills with high production rates where each additional day of campaign life has significant revenue impact, and in operations where gunning material costs are a visible budget line that can be directly reduced through targeted application. A detailed ROI model should account for your specific relining cost, gunning consumption rates, and production value per heat to project a realistic payback period for your operation.
Book a demo and request an ROI model built for your vessel parameters.
Stop Guessing on Lining Thickness
Give Your Refractory Team Per-Heat Data Instead of Periodic Snapshots
See how iFactory connects BOF camera and laser data to wear mapping, gunning triggers, and campaign tracking in one integrated system.
