Prediction And Analysis Of Injection Molding Process Parameters Based On Efficacy Coefficient Method And Hopfield

With the development of science and technology,China's manned spaceflight has made remarkable achievements.As an important part of manned spaceflight,spacesuit can effectively protect the environmental factors such as high and low temperature,solar radiation and meteor.As an important part of the space suit,the space mask not only protects the life of the astronauts,but also as a transparent material,which can guarantee the needs of the astronauts' working ability.But in the actual production process,it is difficult to use the traditional injection experience and try to achieve the ideal situation,not only waste raw materials,but also the long test cycle,can not meet the requirements of rapid production.With the development of injection molding CAE technology,software simulation can be used to improve the process conditions,process parameters and other aspects,early detection of defects,and production of more ideal products.This paper takes the space mask as the research object,uses the Lexan PC2730 produced by GE company of the United States to make production.Through the analysis of the actual use process,the optimization target is warpage,volume shrinkage and shear stress.Through the analysis and comparison of the process parameters in the CAE software moldflow,the material temperature,mould temperature,pressure holding time,pressure holding pressure,cooling time,injection time are studied.The effect of factors such as Inter and injection pressure on the injection molding process,the experimental data of the optimized targets are obtained by orthogonal test,and the initial optimum process parameters are obtained by the extremum and variance of the efficiency coefficient method.Then the prediction and simulation are carried out through the hopfield neural network,and the best process parameters are obtained.The specific research contents are as follows:First,due to the need to analyze the model,a good three-dimensional model isIV introduced into the Moldflow software.Through a series of pre-processing,8 different gating systems are determined.Through the analysis and comparison of the 8 design schemes of the gating system,the appropriate gating system is determined,and the suitable process parameters are selected and the space surface is applied to the space surface.The initial results of the cover were analyzed.Secondly,because of the large number of process parameters,many groups of experiments are needed.Orthogonal test is used to reduce the number of tests.Through the establishment of a good orthogonal table,L32(49),or 32 tests,can be used to observe 9 factors and each factor is 4.Substituting the established form into Moldflow software,get the optimal target value under different parameters.Then,because of the more optimization objectives,it is impossible to choose a set of suitable process parameters to optimize,using the efficiency coefficient method to calculate simple and high reliability.The optimization target value under different parameters is replaced by the formula of efficiency coefficient method.Because of the different physical quantity of the optimized target,the influence degree of each index to the product is also different.In order to determine the weight of each index,first of all,it is necessary to do no dimensionless and quantitative treatment of the optimized targets.In this paper,the extreme value method is used to do no dimensionless treatment of the optimized targets,and the weight of the optimized targets is determined by the standard deviation method,and the calculated weight is replaced by the method of efficiency coefficient.Step the best parameter.Finally,it is necessary to select a group of optimal process parameters in a number of experiments.The Hopfield neural network can be used to predict the network,reduce the number of test times,classify the evaluation indexes corresponding to the process parameters,and take the values of each evaluation value as the ideal evaluation index of each grade,and get the ideal coding of grade evaluation index..And then through four groups of experiments to learn,through the Hopfield neural network to get the simulation results,get the best of the third groups,followed by the first group of experiments,then fourth groups of experiments,and finally second groups of experiments.By taking the process parameters into the CAE,the four groups of tests are obtained by the efficiency coefficient method,and the results are sorted to get the best of the third groups,followed by the first group of experiments,then the fourthgroups,and the final second groups,and the corresponding models of the Hopfield neural network training.The type is the same.It shows that the Hopfield neural network is feasible.The Hopfield neural network can be applied to the injection process parameters,and the optimal process parameters are compared,thus the quality of production can be improved.