Process Supervision of Long Products Hot Rolling Mill. Attack or Failure Identification.

Abstract

The aim of this paper is to explore the potential capabilities of nonlinear projectors to represent the integrated behaviour of products when manufacturing involves several plants inside a typical facility of heavy industry. The interest is focused in the integrated supervision of production able to identify sudden failures or infrastructure attacks discovered from the process perspective. In this analysis, not only the feasibility but also computational capabilities are going to be explored. The selected methodology is the case analysis with action research, and the selected facility is a hot rolling mill for long steel products, which involves more than twenty plants or equipment units. The challenge is to define a monitoring procedure able to assess the product as well as the individual behaviour of each equipment. The results obtained allow to define an automatic procedure applicable in intraprocess time, which can have as an additional benefit its visual interpretation from the operator point of view.