Font Size: a A A

Plate Rolling Process Control Model And Simulation Study

Posted on:2009-05-27Degree:MasterType:Thesis
Country:ChinaCandidate:H L WangFull Text:PDF
GTID:2191360302975741Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
For process control of modern medium plate mill is very complex, traditional rolling control model can't adapt modernized control requirement. And improving further precision of prediction models based on BP Neural network is limited because of deficiencies of BP arithmetic. So the medium plate mill control models are constructed by use of more effective RBF Neural network in the paper. In the meantime, thermo-mechanical coupled FEM is used for simulation research of the medium plate mill mechanics behavior, and numerical simulation method of rolling process is further perfect.The main influencing factors of medium plate rolling process are analyzed. On the basis of the data obtained from production field in 4200 medium plate mill, mechanics parameter models are constructed by use of RBF network, including influential coefficient of stressed state, flow stress of carbon steel, rolling load. For improving precision and rapidity of convergence of the model, related network model parameters are researched. By comparing three models, it is obtained that prediction precision of network model will reduce with increment of parameters of network input layer. And compared with prediction results of BP network, the result indicates that RBF network is more stable and has better accuracy and speed of training.According to the rolling rules of some manufactory, the 3D FEA model of rolling process is developed, which considers the feature of medium plate rolling, material and contact nonlinearity, inter-relationship of stress, strain and temperature of rolled piece with the material alloyed steel, and treats the load and boundary condition, and interrelationship of temperature and deforming field. Three-dimensional thermo-mechanical coupled FE model of rolling process is established based on LS-DYNA3D finite element program. Through the numerical simulation of rolling, the plastic deformation rule and the changes of stress, strain, temperature and rolling force is described. And compared with the data obtained from production field, the result indicates that thermo-mechanical coupling simulation of rolling process based on LS-DYNA finite element program is feasible.Finally, nonlinear parametric excitation vibration characters of 4200 medium plate mill are researched by the numeric simulation method, and bifurcation diagram and Poincare sections are plotted. Phenomenon of quasi-periodic and chaotic vibrations of the system with vibration frequency is analyzed. By the results of analysis, with increment of rolled piece stiffness, vibration behaviour of the mill system is more complex and the system is unstable more easily.
Keywords/Search Tags:Medium plate, Numerical simulation, RBF neural networks, Model, Nonlinear vibration
PDF Full Text Request
Related items