Research On Multi-objective Optimization Of Automobile Bumper Based On CAE Technology

With the development of economic,injection molding products are widely used in a variety of areas such as electronics, aerospace, automobiles while at the same time people are increasingly concerned about the quality of the products. The injection molding process is complex, and quality of the products is affected by many factors, such as materials, mold, process parameters. The process parameters affect the flow state of melting and have direct impact on the quality of the products. Therefore it is of great necessity to optimize the process parameters in order to get better products.In this paper CAE technology was taken to simulate the forming process of a truck bumper, and then analysed the warpage and volume shrinkage to obtain the optimal parameters. The main contents included are as follows:Firstly, the finite element model of an automobile bumper was created by Moldflow software. Then the impacts of five different processing parameters on warpage and volume shrinkage are investigated based on orthogonal experiment design and numerical simulation. It is concluded that the greater effect of processing parameters on warpage are in the order of packing time, packing pressure, injection time, melt temperature, and mold temperature accordingly. Parameters such as melting temperature, injection time, packing time, mold temperature and packing pressure had greater influence on the volume shrinkage in turn by means of variance analysis. Optimal parameters for warpage and volume shrinkgae are obtained the based on the orthogonal experiment analysis.Secondly, combined orthogonal test with grey correlation analysis, the melt temperature, mold temperature, injection time, packing time, and packing pressure were selected for the experimental variable, and the warpage and volumetric shrinkage as the quality atandard. Grey correlation is taken as evaluation factors to calculate the average gray correlation degree of different parameters under different values. The higher the value of grey correlation is, the better comprehensive quality of this lever of this parameter can be obtained. And the better parameters combination was selected for both warpage and volumetric shrinkage based on the value of the average gray relational degree. Through the Moldflow simulation the value of warpage obtained increased by0.245mm compared to the minimum value of the orthogonal experiment, which was not large, but the value of volumetric shrinkage decreased to11.98%under the optimized parameters, therefore improving the comprhensive quality of the products.Finally, an artificial network model based RBF algorithm and support vector machine (SVM) model was set up, which can meet the accuracy requirements. From analysis of test samples it is concluded that the model of RBF neural network gets the absolute percentage error within10%, and the model of SVM gets the absolute percentage error within2%after the training samples were trained. It is concluded that SVM model has better fitting and accuracy. And it’s more suitable in dealing with the small sample datas. Moreover, combined SVM model, orthogonal test and grey correlation analysis, the parameters were further optimized and simulated by Moldflow. The results showed that the warpage and the volume shrinkage value were further reduced.The case of bumper indicates that this method is helpful in improving the efficiency and quality of plastic products. And the trained model of SVM can replace Moldflow software to predict experimental results. It can save plenty of time to optimize parameters and accordingly improve the efficiency of the development, and it is of certain significance for the actual production.