The Defect Analysis And Parameter Optimization Of Injection Molding Process With Moldflow

As an efficient parts forming method, injection molding process has widely used in plastic industry for production. The quality of parts can be described in several aspects like the shape, size, appearance, strength, durability, to list a few. Corresponding defects have flash, air cavitation, shrinkage, warpage, crack, welding line, among others. Major factors on formation of defects include materials, process parameters, injection molding equipment, and production environment, Of which process parameters are relatively easy to change, and have a direct impact on melt filling, pressure maintaining, and cooling. Clearly, we can use the CAE software to analyze the injection molding defects and determine the location and severity of the defects. Using the analytical results to improve the injection molding process parameters to obtain high quality of injection molded parts at a relatively low expense has the practical significance.In this paper a plastic part was analyzed with Moldflow, improved the grid quality by decreasing the grid aspect ratio and refining the quality of grid matching rate. Warpage values was set as reference for the convergence of finite element analysis results to get an efficient meshing taking account of the computing resources consumption and the meshing analysis precision. Using Taguchi test to analyze the relationship between shrinkage rate, mark depth, the maximum warpage and process parameters by establishing orthogonal table to assign the test factors and levels, and then set a range analysis of the experimental results., To get a more accurate mapping of the process parameters and synthetical quality, Training a BP network with results data then we can get a better optimization.The multiple objective orthogonal experiment was conducted by Moldflow, sorting the result of synthtical quality by the range analysis, the optimal process parameter combination shows plastic part of the best synthetical quality reaches 71.354. Searching for the best parameters arrangement with the simulated annealing algorithm in network, a better synthetical quality reaches 76.862, Increased by 7.7% compared with the orthogonal experiment.