Application of Rolling Force Learning Mathematic Model for Plate Mill

The forecasting precision of conventional rolling force model in practice is not high,so a new rolling force learning model based on feedback data from the instrument in rolling process is developed.It is very necessary to study the method for division of learning coefficient layers according to the steel type,exit thickness,starting temperature,target thickness and width.Also,the gist and the effectiveness for the method are analysed.The result of on-line application indicates that rolling force model has good precision of prediction with an error less than 3% in last pass and 5% in other passes.