Study And Application Of Design Optimization For Injection Molding

The injection molding is a popular technique of polymer processing, by which multiple parts with tight tolerances and complex shape can be produced in a single operation. Due to its success in delivering high-value added commercial products with high quality, the global competition in the field is intense, which leads to the requirement of high quality standard, shorter producing time and lower operating cost.This dissertation focuses on developing a fast flow simulation methodology and design optimization theory for injection molding. The optimization algorithms, fast flow simulation and numerical simulation technology are integrated to optimize part design and mold design. The relationship between processing condition and filling pattern of single and multiple cavities are studied. The main work are as follow:1. Through studying the flow behavior of melt in cavity and runner system, a fast simulation theory is developed based on reasonable assumption and simplification. Equivalent flow length concept is introduced, by which the melt-front tracing can be transferred to the computation of the shortest path on part. Two short shot experiments demonstrated the efficiency and accuracy of the proposed theory in predicting the filling pattern and the location of weld lines and air trap. For a model with 6065 nodes and 11700 elements, it needs only 128s in Pentium IV 450 machine, while MPI (numerical simulation software) needs 1845s.2. An O(|e|)-time algorithm is proposed for calculating the approximate shortest pathon finite element mesh model which can be represented by a weighted graph. This method combines FSPA(Fast Shortest Path Algorithm), used to dynamically calculate shortest path on weighted graph, and edge subdivision technology. By iteratively subdividing relate triangle edges adjoined the shortest path and constructing new subgraph, a shortest path with high approximation accuracy can be achieved. The method is employed successfully to tracing melt front of injection molding. It can also be applied to robotics, GIS, route finding, and so on.3. The effect of variation of processing condition to filling pattern for single and multiple cavities is studied by numerical simulation and short shot experiment respectively. It is found that processing condition has little effect on filling pattern for single cavity mold with uniform thickness. For single cavity mold with large thickness difference or multi cavities mold, it may change the entire filling pattern when the variation of processing condition cause flow hesitation.4. An optimization model of part thickness in injection molding is introduced, in which the aim is to get balanced filling pattern. The idea of steepest descent methodintegrated with fast simulation is used to minimize the arithmetical average difference between equivalent flow length of boundary nodes and reference value, and an optimal part thickness distribution can be obtained. Based on the result, the number of part regions can be determined. Then modified complex method combined with numerical simulation is applied to further searching optimal part region thickness and achieve uniform filling pattern by minimizing the arithmetical average time difference between the time reaching to boundary nodes and reference time. The examples demonstrate that the proposed method is viable and robust, and the unbalanced percent can be drop to 5%.5. For mold with limited gate design space and high quality requirement, a methodology, which combines modified hill-climbing algorithm and numerical simulation, is used to optimize gate location to achieve minimum warpage and improve material performance.6. On the basis of fast simulation, a new gate optimization algorithm for mold with multi-gates or big feasible design space is proposed. Exhausted method and adaptive simulated annealing genetic algorithm are used to optimize single gate and multi-gate mold respectively. The models analyzed demonstrate that the proposed method is promising and the computation time is satisfied.