The Molding Injection Quality Prediction Of Appliances' Shell Based On BP Neural Network And Ant Colony Algorithm

Plastics and cement, metal, wood were commonly known as the four national economy materials. Plastics play an important part in the growth of the national economy. With many advantages as excellent mechanical properties, processing properties, light weight, corrosion resistance, electrical insulation performance, high specific strength and so on, plastics have a wide range of applications in various fields. In order to reduce weight and cost, packaging, automotive, and household appliances industries mostly use plastic parts to replace metal parts. Only through the use of molding can plastic be a useful commodity. Injection molding process is an important forming method. According to statistics, the vast majority of plastic parts were got through injection molding process, which is the most important forming method.Recently, the maturity of the finite element theory and computer technology made numerical simulation technology which based on the finite element theory gradually been promoted and applied. Numerical simulation technolog also provides a good tool for the study and control of defects in injection molding. Combined numerical simulation technolog with numerical optimization method is a new way to the quality control for injection molding.Under the background of the above-mentioned technology, this paper has done research on the controlling method on wrapage, which is the main defect of molding. Put forward the necessity to research appliance shell forming by analyzing development situation of the household appliances industry. Using LCD TV shell as the object, uniform design experimental data as the basis, MoldFlow injection molding simulation as a tool, injection process parameters like injection temperature, mold temperature, injection time, packing time, cooling time as experimental factors, the practical using levels of the experimental factors as the level, the warpage of injection molding simulation results as the objective parameter to complete the construction of the uniform table. Then done data analysis by SPSS software to study the importance level of parameters on the wrapage result, and construct the regression model. Then create warpage model based on BP neural networks. Since BP algorithm is defective, we construct a new method with BP neural network and ant colony algrithm, which combines the global performance, heuristic advantages of ant colony algorithm and generalization performance of neural network. Applied this menhod to the forecasting quality of injection molding, can significantly improve forecast accuracy.The characters of this paper consist in the following two factors:1) A detailed data analysis by SPSS to determine the importance level of various factors, and the creation of the regression equation. 2) By introducing ant colony algorithm to optimize weights and the threshold of the neural network, we get a new method. Using the new method to construct new model can greatly improved prediction accuracy, which offers a new way to solve the problem, and has an important application value. And also give a reference for the subsequent research, and has a certain academic value.