Multi-objective Optimization And Robust Design Of Sheet Metal Forming Based On Dynamic GABP Model

Sheet metal forming is widely used in automobile manufacturing and aviation manufacturing.In sheet metal forming optimization,the forming quality of the part can be significantly improved by optimizing the forming process parameters,die size and friction conditions.In addition,aiming at the problem of component forming quality fluctuation under the optimal process conditions,the probability of part failure can be obviously reduced through the robust design and the forming quality stability can be improved.This paper briefly describes the research background of sheet metal forming optimization and robust design,and reviews the research status of sheet metal forming optimization and robust design.On this basis,the research on sheet metal forming optimization and robust design based on dynamic GABP agent model is carried out.The main contents of this paper are as follows:First of all,combining the experimental design method,BP neural network and genetic algorithm to optimize the connection weights and thresholds of fixed network structure with the aim of minimum mean square error estimation,a construction method of GABP agent model is proposed.On this basis,in order to reduce the number of function evaluation,reduce the optimization cost and improve the efficiency of optimization,a dynamic update strategy of GABP model is proposed based on Latin hypercube sampling and maximum and minimum distance order plus some criteria.This strategy ensures that the approximate accuracy of the model in the global region and the local optimal region gradually increases until convergence.Through the typical mathematical test function,the validity and efficiency of the construction method of GABP agent model and its dynamic update strategy are verified.Secondly,based on the dynamic GABP agent model method and gray relational decision-making,a multi-objective optimization design of box-shaped parts for deep drawing forming quality reduction and creasing is carried out.With the blank holder force,punch fillet radius and friction coefficient as the design variables,a proxy model with high prediction accuracy was established based on the finite element analysis using the test design samples.Multi-objective optimization is transformed into single-objective optimization using gray relational decision-making,and a multi-objective optimization solution process of box-shaped part forming is established.The optimization results show that,compared with the traditional static proxy model method,the dynamic GABP proxy model method can reduce the number of finite element calls and improve the optimization efficiency.In addition,the multi-objective processing method based on gray relational decision making can simplify the optimization process and obtain an effective coordination solution of thinning and wrinkling,and improve the forming quality of the square box.Thirdly,in order to solve the problem of sheet metal forming quality fluctuation under the traditional optimal process conditions,a multi-objective robust design method of sheet metal forming based on Dual-GABP model and KS function is proposed.According to the range of design variables,the experimental samples were obtained by using Latin hypercube sampling and finite element simulation.The first GABP model of forming quality and design variables was established,and the high precision prediction of forming quality was achieved.On this basis,according to the fluctuation range of design variables,Monte Carlo simulation is used to obtain the mean and standard deviation of the forming quality index,and the second-order GABP model of the forming quality indicator mean and standard deviation is established.The KS function is introduced to deal with the multi-objective optimization problem,which avoids the difficulty of selecting multi-objective optimization weight coefficients.The results of multi-objective robust design of square box parts show that the mean of thinning index increased by 2.26%,the standard deviation decreased by 41.51%,the mean of wrinkling index increased by 7.23% and the standard deviation decreased by 18.05%.In addition to the average value of forming quality index increased slightly,the standard deviation were significantly reduced,indicating a significant reduction in forming quality fluctuations,increased robustness.