Model predictive control (MPC) of injection molding machines

The problem of on-line control of plastic injection molding machines is considered. The specific problem studied is accurate regulation of melt temperature which has a significant effect on quality of the molded parts. A robust system identification and control mothodology is developed which uses canonical variates analysis for identification and model predictive control for regulation. Data collected from an operating machine are used for identification of control system models in two operating states. Temperature regulation performance is analyzed using simulations in which significant disturbances in machine variables and noise are introduced. The results indicate that the proposed methodology is superior to current PID controllers and can be easily implemented in current digital process control computers.