Model Prediction & Iterative Learning Compound Control For Injection Molding Machines

Injection molding is batch-production process that following sequential operation steps. The products are widely used in several domains such as industry, agriculture, construction, packaging etc. With the world energy sources crisis and the excellent capability for plastic, it has a great demand for the plastic parts both the quantity and quality, more and more attention are paid on injection molding. Manual operation dominates in traditional injection molding processes, so the corresponding automation is generally lack. As a result, advanced control strategy and optimal control method are urgently required to improve productive efficiency and quality for plastic product.During injection molding process, injection speed is an important factor to control performance. Therefore, effective control of injection speed is very important by means of different methods. Because non-linearity, time-varying, disturbance, non-accurate mathematic model etc. exist in injection process, PID control method can not do as well as expected, but model predictive control has special advantage. Besides, the repetitive operation in injection process coincides with the application feature of iterative learning control.Considered the facts above comprehensively, a hybrid algorithm of model predictive control and iterative learning control is given based on the combination and improvement of model predictive control and iterative learning control, and the controller is designed also. Furthermore, an iterative learning for predictive time length is proposed which is tuned based on system errors of previous control periods online. Finally, simulation is carried out for the ram velocity control in injection molding process, and the results show that the algorithm is effectiveness and its control performance is better than PID feedback and feed-forward iterative learning control.