Real-Time Melt Temperature Prediction for a GCC Steel Producer
Challenges
- Electric arc furnace operations face significant variability driven by raw material quality, furnace behavior and operator-dependent decisions.
- Operators must decide when to take temperature measurements and when to tap, often on experience rather than predictive insight.
- The result is high variability in heat duration, a risk of overheating or underheating, inefficient energy and electrode use, and lost capacity downstream.
Approach
We developed an AI assistant for Electric Arc Furnace operations to help steel plant operators make better real-time decisions during each heat.
The system predicts melt temperature, recommends the optimal first measurement moment, and supports tapping readiness decisions.
Key features
- Real-time melt temperature prediction
- First-measurement optimizer based on actual furnace behavior
- Process awareness dashboard with live furnace parameters
- Tapping readiness support
- Visualization of energy, charged DRI, specific energy, and process dynamics
Impact
1.5%
Reduction in electrode consumption
up to 1%
Reduction in electricity consumption per ton of steel
Per heat
Decision support during every heat, not after the shift