Closed-loop AI optimisation

Printed 23 December 2022

A brand new wave of synthetic intelligence (AI) techniques are rapidly changing legacy cement course of controls with self-learning optimisation and complicated AI fashions. They enhance high quality whereas permitting mills and kilns to realize unprecedented throughput and vitality financial savings. The closed-loop optimisation of the Roanoke plant in Virginia, USA, demonstrates how swiftly these enhancements might be made. By Othon Manis, TITAN Cement, Greece, and Javier A García and Iñigo Llopis, OPTIMITIVE, Spain.

TITAN’s Roanoke cement plant in Virginia, USA

The digital transformation is revolutionising cement manufacturing. Many cement producers are already implementing digitalisation initiatives or are planning them. Synthetic intelligence (AI) adoption in processes involving advanced determination making is among the most profitable initiatives, with clear returns on funding. That is the case of AI-powered superior management and real-time optimisation, the place property just like the uncooked mill, kiln and ending mills are outfitted with AI software program, which is related to their management system, to use optimum setpoints in closed loop, totally operated by AI. Such examples, which sounded nearly not possible only some years in the past, have gotten a actuality in a rising variety of cement vegetation the world over and are yielding spectacular returns of funding in lower than a yr generally.
These extremely environment friendly and versatile good techniques are changing legacy superior controls, based mostly on decades-old applied sciences like fuzzy management or linear MPC. AI outperforms fuzzy management simply as a result of the setpoints that the latter supplies reply to condition-action guidelines that don’t depend on an optimisation algorithm.

To proceed studying this story and get entry to all Information, Articles and Video sections of the CemNet.com web site, please Register for a subscription to Worldwide Cement Overview or Login