| Automotive cover panel forming is a key link in the process of car body manufacturing, which has high requirement in the size and shape accuracy. Moreover, the parts are required to have smooth surface, uniform corner and clear decorative feature line. However, due to the complicated and irregular surface as well as the large size of contour, quality defects such as crack, wrinkle and spring-back are prone to occur in the process of drawing forming.Drawing process is the foundation of all body panels stamping operation. Drawing quality of body panel has a great influence on the whole process scheme, which directly determines whether the appearance quality of the product meets the requirements of quality standards. In this paper, a body front fender has been taken as the research object. Stamping forming finite element model has been established and drawing forming process simulated using professional sheet metal forming finite element software Pam stamp. As a result, the quality defects of stamping forming in software through the post-processing module can be observed to analyze the influencing factors of stamping forming defects, stress and strain distribution can be obtained during and after deformation. Therefore, the cycle of mold design obviously can be reduced.Due to the complex process of front fender plastic deformation, the tendency to generate a variety of quality defects and the problem of difficult to determine optimal forming parameters, the grey relational analysis method will be applied to optimize multi-objective process parameters in front fender drawing forming. Aiming to comprehensively measure the effects of blank holder force, friction factor and draw-bead geometry parameters on drawing forming quality and determine the optimal combinations of drawing process parameters, it is necessary to calculate the grey relational coefficient of single-objective function, relational grade of multi-objective function and average grey relational grade of objective function through the simulation data. The key factors based on grey relational analysis optimization are 900 KN of BHF, 5 mm of convex draw-bead radius, 4 mm of concave draw-bead radius, 5 mm of convex draw-bead height, 12 mm of concave draw-bead width and 0.12 of friction coefficient. Numerical simulation experiment and product verification show that the quality of parts significantly improves under this optimized process parameters. It indicates that the application of grey relational analysis method in optimizing multi-objective process parameters of front fender drawing forming is effective and feasible, which can provides effective guidance for optimizing the stamping process and designing die structure.Finally, combining the BP neural network theory, the BP network forming quality forecast program has been written based on Matlab platform. Then, complex nonlinear mapping relationship between sheet metal forming quality and forming process parameters can be constructed through this program. With orthogonal test data as sample data, mean square error has been calculated under the condition of different training times and different number of hidden layer nodes, which can determine the optimal structure of BP network is 6-9-2. Furthermore, prediction model of front fender stamping forming between quality and process parameters will be established. By changing a process parameter value and keeping other parameters constant, a given value was input making use of the trained BP network prediction model and the corresponding output was compared with the simulation results, which can detect the accuracy of the prediction data and further analyze the influence of changes in process parameters on stamping quality. Simulation time and product development cycle obvious can be shorted. |
PDF Full Text Request