Subsection cooling setup and application in ultra wide hot-rolled aluminum strips based on multiple prediction

The setup of subsection cooling is usually based on the experience of technicians and operators, lack of theoretical basis. So, a two dimensional temperature field model of work roll was established using the alternating difference method. Based on the prediction of the change of thermal contours during threading, an iteration algorithm was developed to get the basic pattern for the setup model, under which the thermal contour could be maintained. RBF neural network was established to predict the strip profile by giving quadratic and quartic profiles, and with fuzzy inference, adjustment of the basic pattern was determined based on the predicted profile error. Combining the prediction of thermal model and profile model, the final setup pattern was obtained, under which the thermal contour can be effectively controlled to reduce the profile error during threading. © All right reserved.