AI Plant Control for Steel-works utilities, Germany & Switzerland
3%
Increase in product yield (Argon recovery)
2%
Reduction in Energy Consumption
60-80%
Reduced effort to implement and maintain model

Annual cost savings

Business Challenge
Operators need to optimize product yield to efficiently meet demand. A potential Linear Model Predictive Control (LMPC) solution should be deployed quickly, with continuous model updates requiring minimal effort.

Solutions Deployed
AI Plant Control (AIPC)
- Robust operations across operational modes with different machines in operation
- Model extension for gas case: two weeks only

Benefits
- 3% increase in argon recovery, significantly increasing plant productivity.
- 2% reduction in energy consumption, with the potential for more saving over time as model learns and improves its performance.
- 60-80% reduced effort to implement & maintain model, compared to non-AI driven LMPC models.
These improvements collectively delivered annual savings of 300k USD