Online adaptation model for accelerated cooling process in plate mill

Abstract

Two online adaptation models tuning the flowrate of cooling water in plate mill are developed based on the feedback and feedforward control algorithm. In the feedback control based model, a multiplication factor is adopted which is composed of three adaptation coefficients. The calculations of the three coefficients are designed taking process variations into account. After tested on an offline simulator, the adaptation model is installed on the online process resulting in 10% increase of the accuracy rate of the final cooling temperature. And a neural network model is developed to factor in the variations of the prior processes. It uses 24 process variables as inputs and predicts the final cooling temperature. Comparing with the measured data, the predicted temperatures show an accuracy of ±15 °C.