Self-optimizing MPC of melt temperature in injection moulding

Area of Science:

• Polymer processing
• Control engineering
• Manufacturing technology

Background:

• Plastic injection molding parameters are nonlinear and interconnected.
• Precise control of melt temperature is crucial for product quality, reducing setup time, and preventing thermal degradation.
• Existing control methods struggle with the complex dynamics of injection molding machines.

Purpose of the Study:

• To develop and implement an advanced control strategy for precise plastic melt temperature management.
• To create a generic, self-optimizing model predictive control (MPC) methodology adaptable to various polymers and machines.
• To minimize overshoot, material degradation, and operator setup time in injection molding.

Main Methods:

• Conducted step response testing on industrial injection molding machine (IMM) barrel heating zones.
• Developed a multiple-input-multiple-output (MIMO) model predictive control (MPC) strategy based on experimental data.
• Implemented a learning and self-optimizing MPC methodology for adaptable melt temperature control.

Main Results:

• Experimental step responses revealed significant process coupling between heating zones.
• The implemented MPC strategy achieved good melt temperature control with negligible oscillations.
• The system demonstrated effective control across varying setpoint trajectories and different machine configurations.

Conclusions:

• The developed MPC methodology provides robust and adaptable melt temperature control for plastic injection molding.
• Negligible overshoot and oscillations ensure material integrity and reduce production inefficiencies.
• This approach enhances product consistency, minimizes waste, and significantly reduces machine operator setup time.