Study On Flatness Control For Foil And Thin Strip Cold Rolling Process

Based on flatness control project of aluminum foil and four-stand cold strip rolling mill(TCM) in Benxi Iron & Steel Corporation. The flatness control systems of foil rolling mill and strip rolling mill have been thoroughly investigated. Aimed at problems that flatness actuator efficiency of foil and strip are set up upon width of foil and strip, which does not result in eliminating flatness error maximally. The kernel parameters and key model of multivariable optimization flatness control system, namely flatness actuator efficiency and multivariable optimization algorithm, were analyzed in detail. Finite element model for calculating foil and thin strip flatness actuator efficiency were constructed to investigate factors such as strip width, rolling force, bending force and tilting force et al on flatness actuator efficiency. Duel method is used to develop multivariable optimization algorithm program in C language. Closed-loop feedback control module is developed. The main works are presented as follows:(1) Traditional cold rolling theory in which the roll profile is treated as arc is successfully applied to strip cold rolling process; however, huge error appears for foil cold rolling process. A 2D finite element model for cold foil rolling process was established by utilizing ABAQUS/Explicit commercial software package to calculate contact stress distribution and roll profile in roll bite. It was found that a neutral zone where the thickness of foil did not change in roll bite,which is consistent with "mattress model" of FLECK et al theory for cold foil rolling.(2) In order to further optimize bending force preset, Finite element model for presetting bending force of thin strip was established. The profiles of roll gap under load and thin strip ware decomposed by Chebyshev polynomia. The model was got according to profile match between roll gap under load and thin strip, namely the equivalence of Chebyshev quadratic coefficient.(3) Determing kernel parameters and key model in flatness control system. VAI flatness control system of 4-high foil rolling including setup of target flatness curve.flatness feedforward control,flatness feedbackward control and spraying control were investigated thoroughly as well as ABB flatness control system of 4-high thin strip rolling including setup of target flatness curve.flatness feedforward control,flatness feedbackward control. These two flatness control systems were contrasted to determine the kernel parameters and kernel model.(4) Flatness actuator efficiency of foil and strip are set up upon width of foil and strip in original, which does not result in eliminating flatness error maximumly. The model for calculating foil and strip flatness actuator efficiency was established by finite element method. Influencing factors on flatness actuator efficiency such as width, rolling force, bending force and tilting force et al have been analyzed to put forward control strategy of flatness actuator efficiency.(5) Based on deformation characteristics of foil rolling, traditional pattern recognization flatness control in which on degree term and quadratic term of flatness error are eliminated by tilting and bending respectively cannot be applied to foil flatness control effectively. Dual method was used to develop multivariable optimization algorithm program in C language. Multivariable optimization flatness control system gains wide range application and excellent application effect.(6) Aimed at solving flatness problems in practical production, flatness actuator efficiency is adjusted on the basis of multivariable optimization flatness control principle. Feed forward bend gain modifier is changed from 3.7 to 4.0, resulting in 50% flatness error reduction. Bending flatness actuator efficiency scale is modified from 1.0 tO 0.96, following by 6% increase of bending force and 50% reduction of flatness reduction in two measuring zones of drive sied and operator side. Debugging 1 tilting flatness actuator is used, reducing tilting force by 4.04% and flatness error by 60%.