AI Data Analysis and SOM for the Monitoring and Improvement of Quality in Rolled Steel Bars

In the steel sector, within the production of high value hot rolled long products, assessment of components health and compliance with quality targets are fundamental aspects that need to be ensured. In this paper, an AI–based system for monitoring the conditions of the rolling process of bars with round section and estimating the final bars ovality is presented. The developed system, based on a Self Organising Map, is trained and evaluated by using data coming from a real plant and pre–processed by means of advanced AI techniques. The system allows monitoring a wide range of process variables affecting the ovality issue and actively supports plant operators in the task of avoiding quality–critical conditions and possible machine faults.