NZ Steel

Context

NZS Hot Strip Mill Plant (HSM) has the requirements to improve our Finishing Mill Roll Gap Setup to achieve on-gauge for the Head End of bar. The performance of our Gauge hasn’t been good in our head-end.
Currently:-

• NZS HSM has a Finishing Mill Roll Gap Model (from Toshiba) that predicts the gap required for each stand to achieve the head-end strip thickness.
• This Finishing Mill Roll Gap Model will then send a command to setup the gaps in the rolls for each stand in our HSM
• This Finishing Mill Roll Gap Model works well in Short Term Learning (Bar to Bar on same chemistry and product dimensions) but not Long Term Learning.

Challenge

Can a Model be designed to produce an accurate predicted finishing mill roll gap to achieve output thickness of hot strip to be within the specification required by customer.
With this model, reduce the non-prime products due to thickness out of specification.

Requirements

Setup Roadmap to understand the current model and opportunity.
New Roll Gap Model based on all known process variable for achieving accurate on-gauge in Short Term Learning and Long Term Learning.
New Roll Gap Model must be robust with capability:-

• For Process Engineers to fine-tune in future.
• Capable of handling multiple input chemistries, dimensions and running conditions of the plant.
• Ability to be integrated into our current PLC language.

Improvement of non-prime products due to thickness out of specification by 20%.