Optimization of Rollgap Self-learning Algorithm in Tandem Hot Rolled Strip Finishing Mill

The thickness precision is an important indicator in strip production, in which a self-learning with high precise model is necessary. In this paper, tacking new data collection and processing, looper speed compensation and more influencing factors into account, an optimized rollgap self-learning model was proposed. With the help of algorithm optimization of Newton-Raphson method, the calculation accuracy are enhanced, and make the actual thickness more approximate to the target value. The application of a 700mm tandem hot strip rolling mill shows that the model could meet the demands of on-line control with high computing precision, and the thickness accuracy are raised to a higher level.