Research On The Technical Parameters Optimization Of Multi-Point Forming For Sheet Metal Parts

Multi-point forming (MPF) is a flexible manufacturing technology for three-dimensional sheet metal parts. In MPF, the conventional monolithic die is divided into a matrix of discrete elements. Each element's height and velocity can be controlled by computer. The forming process of sheet metal is implemented by the contour surface of the elements groups.During the MPF process, because the contacting condition between elements and sheet metal varies continuously and the load condition is complex, so the dynamic explicit algorithm is the best choice in FEM analysis. In this thesis, dynamic explicit finite element algorithm is used in the simulation of the MPF process.The MPF of sheet metal is prone to appear some formed defects, such as wrinkle and dimple etc. In order to avoid forming defects, we must modify some parameters repeatedly. Based on numerical simulation technology and optimized designed theory, the process optimized design is one of the best methods of material processing. It can promote designed efficiency and quality.In order to optimize the process parameters we need establish object function, but among these parameters there is no obvious function relation. In this thesis, we use Artificial Neural Network(ANN), because it abandons the conventional method. It exercises a lot of discrete datum, distills some domanial information and the information is expressed as magnitude and distributing of network inter nuncial datum, consequently establishes corresponding systemic model. Genetic Algorithms(GA) is a innovative artificial search method. So we use ANN to establish object function model and GA to optimize the process parameters, conbining finite element analysis,ANN and GA. And we use this method to optimize the process parameters. It offers effective means to predict and optimize the multi-point forming.Main works are as following:(1)Analyse the process of MPF for sheet metal parts with the technique of finite element. Establish the elastic and plastic finite element mathematical mouldfor large deformation and big displacement. Go along the numerical simulation aiming at MPF with a quarter box part of 7.5cm.(2)Choose optimized objective, use numerical simulation technical to acquire corresponding objective function value of parameter combination. Use parameter combination as input vector of ANN and object function as output vector to construct ANN model, and utilize the datum distilled from the calculating results of numerical simulation as network study swatch to exercise constructing network. After exercising we could acquire the network of powerful and propagable.(3)Establish an objective function of optimized problem and lacking material expression from the ANN model, acquire objective function value that we need.(4)Combine the ANN consequence program with GA optimized program, optimize the process parameters and acquire the best process parameters value. Utility numerical simulation to validate the optimized result.In this thesis, we utility the optimized method of combination ANN with GA to study the optimization of process parameters primarily in order to build MPF process parameters optimized system and realize intellective optimizing of process parameters.