| Automobile covering forming technology and die design is the key technology that limits the development of car industry. Along with the development of the numerical analysis technology and computer technology, simulation technology gradually has been applied to the automobile cover part stamping mould quality control and the optimization design. But the current application is limited to replace repeated testing molds. At this stage, the artificial intelligence and other advanced technology combined with the numerical simulation method is used to predict the product quality and forming process parameter optimization, and it has become a research tendency.In this paper, based on the theoretical knowledge of numerical simulation of sheet metal technology, plastic forming and artificial neural networks theory and so on, the inside and outside boards of hood are taken as research subject, the drawing process of them is simulated and analyzed with the finite element numerical simulation software DYNAFORM, the important process parameters of the effect on the forming quality are evaluated with the modern design optimization method, a BP neural network to predict forming quality model has been established according to the experimental data. The main research contents are as follows:1.Numerical simulation theory of the research present situation and the development trend are reviewed, the development process of the artificial neural networks is introduced, the key technology of the sheet numerical simulation of deep drawing are analyzed, including the dynamic explicit, elastic-plastic constitutive relation, unit theory, contacts and friction treatment methods and adaptive meshing technique, etc.2. The drawing process of typical cylindrical is simulated by finite element numerical simulation software DYNAFROM, by analyzing FLD and thickness distribution chart, it is known that the main defect of cylinder part is flange area wrinkling. A group of optimized value of blank holder force(BHF) is obtained by setting different BHF value. It is of great application value to guide the actual production of the part with the group of optimized blank holder force value.3. The drawing process of the inside and outside boards of hood is simulated with DYNAFROM software, the main defects of the outside board, including cracking and wrinkling, and the main process parameters which affect the forming defects is analyzed; the influence of blank holder force, punch speed, clearances between punch and die, friction coefficient on forming defects cracking and wrinkling are assessed with the orthogonal experiment method, the optimized value of them are obtained.4. The BP neural network model for the quality prediction of deep drawing is established through the orthogonal experiment data which obtained from the numerical simulation of the outside board drawing process, through a large number of learning samples after training, network and numerical simulation of the predicted value relative error is small, the model is reliable.The research showed that the stress-strain and forming defects in the drawing forming process of inside board of hood could be predicted by using numerical simulation technology. Combined with artificial neural network model in predicting forming defects could reach higher exactness, when it was applied in the process parameters optimization and quality control, its simulation time period is shorter compared with single numerical simulation, which is of great significant to shorten development cycle, improve the production efficiency and further optimization of process parameters. |
PDF Full Text Request