Simulation And Process Parameters Optimization Study For Curved Flanging

This paper selected object of study the surface flange parts for the automotive, including the stretch flange and compression flange. When the production is carried out in accordance with the original factory process, due to the deformation of the material is complex on the portion of the stretchable flanging and thickness distribution lead to in the area on the outer edge of the material will rupture,after modify the mold the factory persists after the high rejection rate, production efficiency is not high, wasting a lot of manpower and material costs.In order to solve the problem of the surface flanging stamping partial rupture, combined with the product structure and forming characteristics proposed flanging curved the preformed composite processing program.The original factory program to simulate the use of finite element analysis technology, compared to the actual defect, numerical simulation results highly consistent physical, subsequently after the first flange of the proposed preforming and surface flanging, bending, bent flanging step forming three traditional forming numerical simulation, select the feature points in the form of contrast under different forming parts key parts of stress and strain changes, indicating that the surface flanging forming process is complicated by the force, by forming way affect the stress-strain changes, these will eventually affect the forming quality.Surface flanging pieces defect on the basis of numerical simulation preformed angle,die radius,friction coefficient,the punch and die gap forming quality influencing factors based on BP neural network curved flange pieces predictable mathematical model.Arranged by orthogonal test program, and the program of numerical simulation. Then arranged for four forming has been simulated using four sets of data to test the network model.Test results comparing the numerical simulation results show that the model is correct, effectiveness, can provide an effective means for the prediction and optimization of surface flanging forming qualityUse BP neural network which after training and genetic algorithm for coupling, set up a fuzzy function relationship about genetic algorithm neural and to optimization,after get each combination of parameters to simulate it. Obtained by simulation parts forming a small amount of thinning, it have no rupture trend,it have positive significance to improve the quality of forming to modify the mold.It also show genetic algorithm coupled neural network algorithm proposed in this paper of preforming surface flanging stamping processing is applied...