Real-time autonomous control of a continuous macroscopic process as demonstrated by plastic forming

Area of Science:

• Materials Science
• Engineering
• Automation

Background:

• Current research and manufacturing demand more adaptable and efficient processes.
• Continuous material processing often involves complex, interdependent control factors.

Purpose of the Study:

• To develop and demonstrate a continuous autonomous system for material processing.
• To autonomously determine process conditions for specified plastic film dimensions.

Main Methods:

• The system integrates real-time, in situ characterization with an active-learning-based decision-making processor.
• Applied to a plastic film forming process, it autonomously adjusted control factors (material supply rate, applied force, viscosity).

Main Results:

• Successfully identified stable process conditions for nine different film dimensions with high accuracy (R²=0.87 for width, R²=0.90 for thickness).
• Achieved an average of 11 characterization-adjustment iterations and 19 minutes processing time per dimension.
• The active learning algorithm enabled exploration with zero initial training data, avoiding common control pitfalls.

Conclusions:

• The developed autonomous system effectively controls continuous plastic film formation without human intervention.
• The system's adaptable design has broad implications for accelerating progress in both research and industrial material processing.
• The core concept is transferable to other continuous material processing systems.