Cross-Phase Model-Based Predictive Cavity Pressure Control in Injection Molding

Sebastian Stemmler1, Muzaffer Ay1, Marko Vukovic2, Dirk Abel3, Julian Heinisch4, Christian Hopmann4

  • 1RWTH Aachen University
  • 2RWTH Aachen University, Institute of Automatic Control
  • 3Aachen University
  • 4Institute of Plastics Processing at RWTH Aachen University

Details

15:30 - 15:50 | Mon 19 Aug | Lau, 6-213 | MoC6.1

Session: Predictive Control 2

Abstract

The injection molding of thermoplastics is one of the most efficient manufacturing processes for autonomous manufacturing of plastic parts. The process is commonly divided into two main phases: the injection phase and the packing phase. During the injection phase the injection velocity is conventionally controlled. In contrast, the packing phase is pressure controlled. Due to the different control objectives within these two phases the control strategy is changed regarding a switch-over point, which leads to issues in process control. Therefore, a cross-phase control strategy is required which avoids the switch-over in order to optimize the process control strategy. In this contribution, a model-based predictive controller (MPC) is presented which considers the process behavior during the injection phase as well as the packing phase. Therefore, a physical-motivated model is developed for both phases. Afterwards, this model is continuously peace-wise linearized around the current operation point in each time instance. Furthermore, the measured signals are filtered and the required system states are estimated by an Extended Kalman Filter (EKF). Then, the presented approach is applied to a servo-electric injection molding machine in order to empirically validate the mentioned approach. The controller shows good tracking performance for both the injection phase and the packing phase.