Packaging steel strip model adaptive characteristics

Packaging steel strip model adaptive characteristics of abstract mathematical model with various assumptions, plus the average properties of modeling environment makes the model precision is restricted, at that time under the environment of the measured data is used to analyse the model adaptive and self learning is to improve the accuracy of model set in a specific environment important method. Flatness, Convex degree) Model involves a complex mechanism of the factors influencing load roller shape, such as thermal roll shape and its changes under subsection cooling, roll wear model as well as related to the rolling force and lateral stiffness, etc. , related to the force of bending roll so complicate with the calculation of loaded roll gap shape. If it can be measured by cross section shape of each frame exit strip thickness (even if it is two o 'clock Center point) and one side of the strip edge landmark Measurements, will make the crown model of adaptive and self learning easier. Most of the current at the end of the cold even Zagreb frame crown instrument has just started in practical application. So in terms of flatness detection instrument lags far behind the thickness control strip shape control, giving flatness model adaptive and self-learning bring great difficulties. The current commonly used & other; Self learning & throughout; Is: when the end of the frame of flatness feedback control use of finished strip shape is improved after the feedback control of roll bending force and roll up move save for self learning, such as to make the volume of steel to improve. As a result of the rolled strip flatness defect is by each frame export strip accumulated relative crown is not identical, ideal method should be the first to ensure the rack exports relative convex degree is equal to the relative crown of hot milk material, and then by the end of the flatness feedback control to fine adjustment. Like alone at the end of the front frame to correct deviation of each frame or remember deposit at the end of the frame after the feedback control of roll bending force, CVC ( Or HC mill) Roll position is likely to fault ( Each frame in front of the relative crown is not identical) Is wrong, Only at the end of the study of feedback control results) 。 If through the accumulation of experience, of course, each frame of the cold rolling roll bending force and roll setting to ensure every rack exports nearly identical relative convex degree or guarantee better strip flatness, between each frame is adopted to remember deposit at the end of the frame after the feedback control of roll bending force and roll values for self learning, to improve strip shape is good. Self-learning and exponential smoothing method is used to model, the self-learning method to model and thickness is similar.