Modelado y optimización de procesos de laminado con transferencia de textura

Abstract

This thesis presents a methodology applied to model and optimize some rolled sheet processes. This line of investigation arise from a research project funded by the European Commission, which aim was to develop a new steel sheet rolled process. This process, called hybrid, is the result of the combination of skin-pass process and tension-leveling process. The process of skin-pass or temper -rolling is a type of cold-rolled sheet which aim to achieve a small reduction in thickness (less than 2%), improved mechanical properties and good surface finish. Furthermore, the tension-leveling is a cold process which involves bending the sheet on either side alternately by means of rollers. The aim is to produce perfectly flat metal sheets homogenizing the longitudinal fibers. The study and optimization of rolling processes is performed by metamodels which combine the finite element method (FEM) techniques and data mining (DM). The FEM is used to model the process by considering the geometry, materials, contacts and working conditions. Then, the results obtained by the model are compared with experimental data to test its validity. Once the model is validated, a set of simulations, which are variations of the same process with different working conditions, is designed in order to generate a database. The FEM can simulate a variety of physical systems with good accuracy but its main disadvantage is the high computation time. That is why DM is used to determine a set of mathematical expressions that relate inputs and outputs process. These expressions are easy to evaluate, allowing the process optimization to get several objectives. In the case studies, these objectives are: to achieve a thickness, maximize roughness transfer and minimize residual stresses. Thus, metamodeling FEM-DM is shown as a valuable methodology for the development and optimization of industrial processes.