AISTech 2017

Sessions Speakers Exhibitors Products Press Releases Show Specials Floor Plan Buzz

Use of Machine Learning to Assess Defects in Hot Rolled Coils (Room 102B)

09 May 17
3:30 PM - 4:00 PM

Tracks: Rolling Process Modeling, Simulation & Control

The North American Stainless hot rolling mill is a highly automated mill that collects and stores into their MES database hundreds of thousands of data points for each coil produced. This paper presents our work using special feature selection and expansion techniques and different machine learning algorithms to generate a model that identifies coils with defects before visual inspection takes place.