Research On Shape Prediction Model Of Hot-rolled Strip Based On Quasi-3D Finite Difference Method

In hot strip rolling,shape defects may cause strip breakages,make subsequent processing difficult and degrade product performance.A fast shape model can not only analyze tandem rolling but also run massive parametric campaigns for industrial process optimization,improving the accuracy of multi-target full-width shape control.However,due to the inherent nonlinearity of plastic deformation and the strong coupling of three-dimensional metal flow,the strip deformation model is the bottleneck in the development of the fast shape model:the calculation time of the finite element model is too long,and the existing fast models require assumptions about the lateral flow pattern or have convergence problem.Therefore,based on the quasi-3D finite difference method,the aim of this research is to establish a strip model which satisfies the combined requirements of accuracy,efficiency and robustness.The main progress is as follows.(1)A rigid-plastic(RP)strip model considering lateral flow is established,which can predict distributions of lateral flow,tension and rolling force.Different from the conventional fast models,RP model does not rely on the assumptions of lateral flow pattern and take the effects of shear stresses into consideration,improving the accuracy.A combined validation based on both the FEM model and the measured profiles in the industrial experiment is carried out;the errors of crown are less than 15%.The modeling strategy—including quasi-3D approximation,decoupling,linearization,discretization and global iterative solution—are found to be effective.RP model takes about 20 ms for a single call,which is suitable for multi-parameter optimization and has the potential for online application.(2)An elasto-viscoplastic(EVP)strip model considering the interstand deformation is established,which can predict the entire evolution of strip shape.Predictions of spread,profile,and residual stress are validated through comparisons with a two-stand finite element model.Compared to the experimental measurements,the error of crown is less than 11%.The calculation time of a seven-stand hot tandem rolling is approximately half a minute,which is two or three orders of magnitude smaller than the cost of FEM model,making multi-pass and multi-case simulation feasible.The evolutionary mechanisms of profile and residual stress are analyzed:during the interstand elastic recovery,the transverse compressive stress releases and the strip velocity tends to be uniform,revealing residual stress after a significant change of stress pattern.The stress relaxation mainly occurs at the edge near the roll bite and therefore increases the edge drop of the profile;it also decreases the center crown by changing the distribution of the rolling pressure in the roll bite.(3)Combining the full-width prediction ability of the RP model and the particle swarm algorithm,the contour and shifting parameters of taper work roll are optimized,a varying-step shifting strategy is proposed to deal with the nonlinear roll ground contour and uneven roll wear.The industrial application verified the new taper roll shifting strategy,as both the edge drop and high spots were alleviated and the rolling length of a schedule was extended by about 10 km,and the wear contour of work roll remains smooth.(4)Using the residual stress prediction of EVP model to analyze the high-order shape defect of stainless steel,the evolution pattern of the high-order residual stress in each stand is obtained,and the relationship between the temperature drop at the edge and the high-order residual stress is revealed.The parameters of the middle variable crown work roll were designed by simulation,and the high-order shape defect was effectively controlled in the industrial test.