Research On The Optimization Of Process Parameters Of The Gas-Assisted Injection Molding Based On Intelligent Algorithm

The gas-assisted injection molding (Abbreviation:GAIM) is a new plasticmolding technology.it has the Characteristics such as the improvement of the productsurface quality, the reduction of the cycle times and the improvement of productionefficiency.In recent years,it has been widely applied in the areas of automotive andhome appliances in Europe, North America, Japan and other developed countries.With the addition of the gas injection phase during the gas assisted injectionmolding process,the molding process is more complicated than the conventionalinjection molding.the molding process relates to a novel process parameters includingthe gas injection pressure, the gas injection time, the melt pre-injection volume, thegas delay time and so on. if the setting of these parameters are unreasonable,the partswill appear the quality problems such as melt short shot, blow and so on.and the gaspenetration is very sensitive to the changing of the process parameters,the smallchanges will lead to the changes of the entire gas penetration state.So how todetermine the critical process parameters affecting the product quality characteristicsindicators,How to determine the affecting trends,the excellent levels and the excellentcombination of these process parameters to the product quality indicators.and how tooptimize the allocation of process parameters to achieve the optimization of themultiple Indicator of the GAIM parts are major problems in the current research.In this paper, using the CAE simulation software MoldFlow as the secondaryplatform, using the air conditioning Portable of a factory as the research object.thispaper studied the influence trends of the process parameters such as the moldtemperature,the melt temperature, the melt pre-injection volume,the gas delay timeand pressure to the product quality indicators such as the gas penetration length,thegas filling percentage,the secondary gas penetration volume and the volume shrinkageof the parts.and then,this paper built the nonlinear relationship between the qualityindicators and various process parameters by using artificial neural network andoptimized the process parameters by using genetic algorithms. The main content ofthis paper was as follows: 1.Using UG established an accurate and reliable CAE simulation model of airconditioner portable.using the CAE analysis software MoldFlow carried on simulationtests. Using the orthogonal and Taugchi expermental design studied the experimentalresults.through the direct analysis,this paper determined the infulence trends of theprocess parameters to the test indicators and the best combination of the processparameters when the quality indicator was best.through the ANOVA method,thispaper determined the weights of each major process parameters and analysised thesignificance of the process parameters to the quality index.Through the gray relationalanalysis,this paper determinde the best results and the corresponding optimalcombination of the process parameters of the multiple indicators.Finally, the optimalcombination was validated.2.Using the analysis data obtained by Taguchi orthogonal experiment as the sampledata,this paper built the predictor model of the quality indicators,which based on thestandards neural network,the cross-validation neural network and the neural networkof genetic algorithm optimization.and got the nonlinear relationship between theinjection molding process and the product quality. finally,through the experimentalvalidation,this paper showed the correctness and feasibility of the predictive models.3. Using the nonlinear mapping function between process parameters and qualityindicators built by the optimized artificial neural network as the fitness function.thispaper established a global process parameter optimization model of the multi-factorsand multiple Indicators based on genetic algorithms.and finally,through theexperimental validation,this paper show the correctness and feasibility of thepredictive models.4.Using VB as a platform,through the automated service technology of theactiveX,this paper constructed the prediction system of Product quality characteristicsincluding the multiple indicator prediction system of the local optimizating neuralnetwork,the multiple indicator prediction system of the neural network improved bythe cross-validation,the multiple indicator prediction system of the neural networkoptimizated by the genetic algorithm.Whichi can not only improve the efficiency ofprogram development,shorten the design cycle,but also improve the usability andadvantages of smart algorithm.