Vimar: waste-free moulding with the TWIST project

The context

Vimar S.p.A. is one of the leading international players in solutions for electrical energy, building automation and home automation, with a global presence in over 100 countries. One of its core processes is the injection moulding of technical polymers for functional and aesthetic components: plates, controls and interfaces that the end customer touches and sees every day. Quality standards are based on the absence of visible defects: the slightest imperfection means the part is scrapped.

The growing variety of product codes and shrinking batch sizes have made production start-up a daily challenge. With no objective data, press set-up relied on the experience of the individual operator: a trial-and-error approach that took numerous iterations and produced a significant amount of scrap before reaching the required quality. There was also a more strategic issue at stake: the know-how of the most experienced technicians was not documented, not transferable, not scalable.

The project

Co-funded by the IRISS call promoted by SMACT and developed in collaboration with the University of Padua, the TWIST project also took shape thanks to Maxfone's technical contribution. The architecture creates a closed loop between press and quality control: the systems developed by Maxfone collect process parameters from the machine in real time while a camera captures every part produced. A machine learning model developed by the University of Padua correlates the two streams, process data and visual quality, generating real-time suggestions for the operator. Training takes place with the experienced operators: every corrective action becomes training data.

The results

During the pilot phase, start-up iterations fell sharply, as did production start-up scrap. Vimar will have a digital body of expertise that does not depend on the presence of any specific operator: the know-how of the most experienced technicians will be codified and accessible to anyone standing in front of a press. Cutting scrap also makes a concrete contribution to sustainability: less polymer wasted and less energy consumed to produce parts destined for disposal.

The next steps include extending monitoring beyond appearance, with dimensional checks and 3D metrology, towards plants capable of continuous learning and autonomous optimisation cycle after cycle.