Identification of multiple operating modes based on fused features for continuous annealing processes

Abstract

This work proposes a new method to identify operating modes for Continuous Annealing Processes (CAP) based on multiple models consisting of key variables with multi-scale features. The existing operating modes in the CAP are first described in detail, allowing key variables influencing the multiple modes to be selected. The distribution characteristics of each variable are then analyzed to select the corresponding detection methods. Furthermore, the multi-scale features of each variable are then fused to establish multiple models for improving the detectability of the process mode. Finally, a case study based on historical data is performed. The proposed method demonstrated identified different modes in CAP and improved identification performances with fused features.