When a 3.8 MTPA integrated steel plant in southeastern India deployed seven quadruped robots from iFactory's Robot Fleet Console in January 2024, the engineering leadership team had a simple objective: inspect the blast furnace cast house, BOF tap floor, and EAF bay every two hours instead of once a week, and see what was being missed. What they found in the first 90 days redefined how the plant thought about equipment reliability. The robot fleet — equipped with thermal cameras, acoustic emission sensors, vibration detection, and gas monitoring — completed 2,184 inspection missions and detected 43 developing equipment failures that had been completely invisible to the existing weekly inspection programme. Those 43 findings were resolved through planned maintenance at a total cost of $127,000. The independent cost estimate for the unplanned failures those 43 issues would have caused: $6.09 million.
How Robotic Inspection Fleet Detected 43 Developing Failures in 90 Days: Steel Plant Case Study
3.8 MTPA plant · 7 quadruped robots · 2,184 missions · 43 failures found · $127K cost vs $6.09M avoided — verified by plant engineering leadership.
Before the Robot Fleet: What Weekly Inspection Was Missing
The plant's human inspection programme was not negligent — it was systematically limited by physics and physiology. At 58°C ambient temperature in the BF cast house, a human inspector could complete one round before heat exposure protocols required a 90-minute rest. The robot had no such constraint. Schedule a robot fleet assessment to see the coverage gap in your plant's current inspection programme.
43 Developing Failures in 90 Days — By Category and Severity
The 43 failures detected were classified by the plant's reliability engineering team into five categories — by failure type, detection sensor, and estimated emergency cost if undetected. Every finding was verified by a maintenance engineer and resolved through planned work before failure occurred.
From Robot Alert to Resolved Work Order — The iFactory Workflow
Every one of the 43 findings followed the same digital workflow — from robot sensor anomaly to CMMS work order to completed planned maintenance — entirely within iFactory's Robot Fleet Console and maintenance management system.
Thermal, vibration, gas, or acoustic sensor reading exceeds AI-calibrated baseline threshold for the specific inspection point. Robot captures multi-sensor data package.
Real-time · <5 secondsiFactory AI classifies the anomaly type — thermal overheating, structural deformation, gas leak, vibration signature — and assigns P1/P2/P3 priority based on failure risk model.
AI processing · <30 secondsAssigned maintenance engineer receives push notification with thermal image, sensor readings, GPS location, equipment ID, and priority classification. Average acknowledgement: 4 minutes.
Mobile push · 4 min avg responseiFactory CMMS auto-generates a planned maintenance work order with all robot data attached — material requirements, skill required, safety permits needed, and SAP PM notification triggered.
Auto-generated · SAP PM syncedMaintenance completes repair during next planned window. Robot re-inspects the same point on next patrol to verify resolution — confirmed thermal return to baseline before work order closure.
Planned window · Robot confirms resolutionCost Verification — $127K Spent vs $6.09M Avoided
The cost avoidance figures were independently estimated by the plant's reliability engineering team using standard industry models for unplanned stoppage cost, emergency repair premium, and cascade damage risk. All 43 planned repairs were verified as complete before the failure threshold was reached.
What the Chief of Reliability Engineering Said
In 90 days the robot fleet found 43 things that our inspection programme had completely missed. The most significant was a developing crack in the BF cast house trough refractory — the robot's thermal camera picked up a 12°C differential that no human would have noticed visually. That single finding alone saved us from a tapping emergency that the maintenance team estimated would have cost $1.8 million to remediate and 4 days of lost production.
Frequently Asked Questions
How was the $6.09M cost avoidance figure calculated?
The plant's reliability engineering team estimated each finding's emergency cost using the plant's own historical data: average unplanned stoppage cost per hour, emergency repair premium factor (3.4× planned), and cascade damage probability from the plant's failure mode register. All estimates were independently reviewed by the plant's finance team.
Did the robot fleet reduce the need for human inspectors?
No human inspectors were eliminated. Instead, they were redeployed from routine patrol to AI-flagged anomaly verification and complex diagnostic tasks — increasing the value of their time while eliminating routine zone entry in high-heat and high-risk areas.
How did iFactory handle the 43 findings through to resolution in the CMMS?
Each anomaly was automatically linked to an iFactory CMMS work order with sensor data attached. Work order priority and scheduling was handled in iFactory, with SAP PM notifications generated simultaneously. Closure required a robot re-inspection confirmation that the thermal or vibration reading had returned to baseline.
What happens after the 90-day period — did the robot fleet continue?
The plant expanded from 7 to 12 robots after Month 3 and signed a 3-year Robot-as-a-Service agreement with iFactory — covering fleet management, AI model updates, and maintenance of the robots themselves as part of the platform subscription.
See Robot Fleet Inspection Live at Your Plant
Demo built around your hazard zones, robot options, and current inspection gaps.
