Optimization for Plastic Injection Molding Operations by Dual Response Surface Method and Non-Linear Programming

• Main Authors: Chi-Wei Hung, 洪啟偉
• Other Authors: Chin-Yin Huang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/86789599922023949304

碩士 === 東海大學 === 工業工程與經營資訊學系 === 96 === In the process of plastic injection molding (PIM), factors of warpage and shrinkage have the notable influence on product quality. Particularly aiming at a small and thin product, the problem that raises warpage and shrinkage will bring into relief more. Molding in the current parameters set, the technicians often need to set or experience to do it using computer simulation software, but such a practice of searching the optimal combination of parameters requires spending a considerable amount of time and cost. For years, the plastic molding methods to improve the quality can be divided into two main categories：the combination of the computer simulation with the optimal technique, among which the Taguchi’s method is the most commonly used in the industry (Rose, 1989), and the response surface method (RSM) based on the statistical methods. The response surface method is a statistical method to solve unknown function with several independent variables and the dependent variables. It combines the regression analysis with the experimental design, and then the optimization of parameters can be found through certain optimization technique. In this thesis, the injection processes of a part were simulated using the commercial software Moldflow. Warpage and shrinkage are considered as two response surfaces in the dual-response surface models (DRS), and then non-linear programming (NLP) is utilized to search the optimal combination of parameters. Nine process parameters including injection time, injection pressure, packing pressure, packing time, cooling time, coolant temperature, mold-open time, melt temperature, and mold surface temperature are considered. Our method of the first phase obtains the better combination of parameters than the Taguchi’s approach. To improve the result of the first phase, the second phase of our method is executed and the result shows that the warpage of the actual value dropped from 0.1559 to 0.1266, and the shrinkage of the actual value dropped from 0.0956 to 0.0808. Lastly, integrating multiple response surfaces of a product’s several points’ quality indices suggests the proposal of the optimization improvement of a whole product.