Research On High-Precision V-Die Air Bending Forming

With the accelerating development of modern industry,the manufacturing industry is demanding higher and higher precision for bent parts.High precision bent parts are crucial for parts assembly,production efficiency and cost control.Due to the structural characteristics of the C-bending machine and the rebound phenomenon after unloading,it leads to the problems of poor angular consistency and large errors between the forming angle and the target angle in the full-length direction of the bent parts,respectively,which cannot meet the high precision requirements of the bent parts.Therefore,the key to improve the accuracy of bent parts is to realize the compensation of deflection in the full-length direction of the bending machine and to establish an accurate rebound prediction model.On the other hand,theoretical analysis is used to study the V-die air bending springback,and the influencing factors of bending springback are analyzed by finite element simulation,and machine learning algorithm is further used to establish the bending springback prediction model and upper die displacement prediction model for most materials.This thesis provides the basis for the construction of an autonomous bending CNC system.The main research work of this thesis is as follows:(1)Based on Ansys Workbench finite element software,study the deflection deformation of bending machine slider and table and the deformation of mechanical compensation device after compensation,and optimize the design of slant wedge angle of mechanical compensation device to realize the bending full-length deflection compensation and improve the bending full-length angle consistency.(2)Based on the plane strain assumption,the theoretical formula of V-die air bending rebound under pure bending conditions is derived using the power exponential hardening elastoplastic material model.Mechanical tensile tests were conducted on Q235 carbon steel plate,304 stainless steel plate and 1060 aluminum plate to obtain material property parameters,and 18 sets of data were obtained from bending experiments on the bending machine for the above three types of plates,and the accuracy of the theoretical rebound formula was evaluated using the bending experimental data.(3)Finite element simulation of V-die air bending rebound based on ABAQUS finite element software was conducted to analyze the effects of sheet thickness,material property parameters(elastic modulus,yield strength,hardening coefficient and hardening index)and die geometry parameters(upper die fillet radius,lower die opening width and lower die fillet radius)on the rebound amount.(4)A Latin hypercube test was designed to combine parametric modeling and Python scripting methods to perform batch simulation of V-die air bending rebound of sheet material in ABAQUS software to obtain 1200 sample data.The DPSO-BPNN rebound prediction model with rebound angle and forming angle as the output and other factors as the input and the DPSO-BPNN upper die displacement prediction model with upper die displacement as the output and other factors as the input were developed by using BP neural network as the research tool.Compared with the bending experimental results,the average errors of forming angle and rebound angle predicted by the DPSO-BPNN model are 0.83° and 0.54°,respectively,and the errors of upper die displacement predicted by the DPSO-BPNN model are within 0.11 mm.