ViscoControl - Control of the injection molding process to compensate for the viscosity fluctuation of recycled plastic
Description
ViscoControl employs Residual Reinforcement Learning to create novel control algorithms to compensate for the viscosity fluctuation of recycled plastic in injection molding. The combination of stability with adaptive, self-learning capabilities will boost the processing of recycled materials.
This leads to improved product quality, reduced material waste, and increased production efficiency. By continuously learning from process data, the system adapts in real time to changing material properties. This makes it especially valuable in sustainable manufacturing environments where material variability is a key challenge.
Key Data
Projectlead
Project team
Etienne Gubler, Dr. Amin Mazloumian
Project partners
Kistler Instrumente AG; Georg Utz Holding AG
Project status
ongoing, started 06/2025
Institute/Centre
Institute of Embedded Systems (InES)
Funding partner
Innosuisse Innovationsprojekt