Research On Key Methods In Numerical Simulation Of Electromagnetic Forming

Electromagnetic forming can improve the forming performance of some materials.The forming process is a high-speed dynamic large-deformation process,which has the characteristics of short forming time and balanced distribution of forming force.The traditional finite element method has been widely utilized by the engineering staffs because of its versatility.However,when dealing with the above problems,there are unavoidable defects such as low accuracy and complicated calculation process.At the same time,it is difficult for numerical algorithms to obtain a good balance between computational cost and computational accuracy when solving large and complex problems.The smooth finite element algorithm uses a low-order unstructured grid,which can obtain high-precision calculation results while ensuring the calculation efficiency.For large metal parts,progressive metal forming will also become the main trend.On this basis,this paper proposes a calculation model of electromagnetic progressive forming based on BP neural network,which combines the BP neural network with the electromagnetic progressive forming process to achieve efficient solution.The specific work of this paper is as follows:1.The gradient weighted finite element method is applied to the numerical calculation of electromagnetic problems of low-order unstructured grids.After the region to be solved is divided into unstructured grids discretely,the gradient-weighted finite element method is calculated by constructing a new smooth domain.This method not only has the advantages of low pre-processing cost and high calculation efficiency of unstructured grids,but also significantly improves the defects of the finite element method due to excessive rigidity and low accuracy.The theoretical derivation and analysis process is extremely simple,which is simple and easy to use,reliable and practical characteristics.By solving and analyzing a series of standard numerical examples and comparing with the results of finite element method and reference analytical solutions,the calculation accuracy,efficiency,convergence and stability of the algorithm in electromagnetic field problems are verified.2.The edge smooth finite element method is used in the numerical calculation of the electromagnetic forming process.Based on the smooth Galerkin weak electromagnetic and mechanical field formulas,a numerical simulation model of the edge smooth finite element method suitable for the electromagnetic forming process is constructed.The effectiveness of this calculation model on electromagnetic forming is verified by analysis of examples such as metal sheet electromagnetic forming.The calculation results show that: based on low-precision units,it can achieve high calculation accuracy and can be applied to complex geometric boundary;strong antigrid distortion ability,and the workpiece structure is deformed severely,and the calculation can be continued;the theoretical method is simple and easy to implement;the calculation efficiency is high.3.The BP neural network is combined with the electromagnetic progressive forming process.In the first forming cycle,it is suggested to introduce grid redivision and grid follow-up technology to the near-field air region in each analysis step based on the coupling mode of time iteration step,the BP neural network can collect and train the first forming results to obtain the force and unique neural network.For the rest of the forming period,the neural network is trained to get the deformation results quickly.It is not necessary to assemble and solve the stiffness matrix in each time step to achieve efficient solution.