| With the injection of mouse as an example, this paper conducts study on injection parameter optimization in plastic injection molding to minimizing product warp by using Moldflow injection molding simulation software as tool. A systematic approach is proposed for injection faults confirmation, off-line parameter optimization, and parameters deviation optimization.First, injection process and injection parameters are introduced in detail. A brief review is made about the optimization techniques for injection parameters. The factors that affect warp in injection molding are analyzed and how Moldflow is used to do injection warp is presented.Secondly, Taguchi experiment design technique is used to optimize injection parameters of warp. When signal-to-noise ratio is used as a measure of quality, the effects of injection parameters on warp fault can be desribed, and the best level combination of different factors can be obtained.Thirdly, Taguchi and Radial Basis Function Artificial Neural Network (RBF) are applied for the optimization of injection molding parameters in sense of product warp. Orthogonal experiment provides training data that give the significance and trend of injection parameters for RBF network. The correlative parameters values are adjusted properly. Then they are inputed to RBF network for off-line warp forecasting. In this way, the injection parameters are optimized gradually.Fourthly, this paper combines Taguchi experiment design, RBF neural network, and fuzzy theory to optimize warp deviation in injection process. RBF neural network with warp deviation as input and injection parameters' deviation as output is employed to produce fuzzy regulation. Warp and injection parameters are defined as correlative fuzzy sets. By using Max-dot fuzzy reasoning inference and ant-fuzzy algorithm, corresponding adjustment values of the injection parameters can be obtained.Finally, conclusions and future study directions are presented. |
PDF Full Text Request