Abstract
In order to obtain the desirable product quality, need to make the optimal process parameter first, but the process parameters numerously and compli- catedly, need to dependent on the experienced operators. Even under the same processing condition, product quality will still be different . In this study, we use the design of experiment in injection molding pro- cess. Because of the injection parameters that we can control are the injection time; vp-switch; packing pressure, injection velocity and the injection pressure, so we choose the L25(56) orthogonal array that in the Taguchi Method. Use the process dynamic data and the part weight that measure from the experiment be the Neural Network training data, we can make a quality predictor that can precisely predict part weight. By the idea of feedback control, we design a single variable process controller. It only adjust the injection parameter which influence the part weight. In the conclusion, the product in this injection molding process can be controlled in the quality requirement and maintain steady operating condition.