Warpage Research Of Thin Shell Plastic Injection Part Based On Data Analysis And Optimization

Warpage is one of the primary product quality defects of thin wall injection molding products. It is caused by the characteristics of polymer material and machining technics,due to the above reason warpage can not be completely avoided. Many factors can affect the thin wall injection product warpage, at the beginning of this paper manufacturing industry intelligent trend had been summarized, based on the thin-walled car wheel trims parts as the research object, the CAE technology and intelligent optimization method were applied to research the effects of the injection molding process parameters on warpage. The factors that affect the warpage were classified from the perspective of uneven shrinkage, and according to the characteristics of controllable factors were researched based on the data analysis. The research is as follows:Main factors causing shrinkage of plastic injection products had been analyzed based on the single factor experiment design method. The result shown process parameters how to affect the product shrinkage separately. The method to control the warpage parts from uneven shrinkage angle had been explored by CAE software. The maximum pressure drop and flow front temperature data were used as an index to determine the plastic plate parts of feasible process parameters and the selection of process parameter optimization range. This method was also used to car wheel trims parts’ s process parameters optimization and simplified wheel eyebrow model size parameters and process parameters optimization.CAE Mathematical model based on the rheological, mechanical, heat transfer analysis was used to be a means of acquiring data. Car wheel trims injection molding process parameters was optimized using the response surface method, the plastic parts of the maximum warpage value decreased from 6.311 mm to 3.367 mm. The thin-walled parts side gate location influence on warpage had been researched. Strip-shaped injection parts’ physical dimension which would be influence the parts warpage had been analyzed using variance analysis by orthogonal test. The effect of different types gate design for quantities of filling balance had been analyzed too.The main geometric feature car wheel trims were extracted to build a simplified model. BP neural network and genetic algorithm were used to analyze and optimize the simplified model geometric parameter and process parameters. geometric parameter,process parameters and maximum part warpage values were input to the BP neural network for training analysis. Through BP neural network hidden layer nodes, using a genetic algorithm of neural network weights and thresholds were optimized. The network training error is reduced to 10-5, training times lower for an average of 104 times, the prediction error also decreases average 2.63%. The use of genetic algorithm to the trained BP neural network prediction model optimization and get maximum part warpage values for 2.238 mm size of the parameters and process parameters combination and is verified by the experiment.