Focus on SMEs experiments: Continuous monitoring of moulding machines using AI at the edge

WHAT ABOUT THE EXPERIMENT? 

​

Within the AI REDGIO 5.0 project, POLYCOM is developing a predictive maintenance system for thermoplastic injection moulding, aiming to prevent production faults and support zero defect manufacturing. The system uses AI algorithms deployed on edge devices to monitor key parameters such as injection pressure, temperature dynamics, and auxiliary systems and identify deviations that may indicate early signs of wear, malfunction, or breakdowns. 

Data is collected continuously from the production setup using standardized protocols and analysed in real time. The monitoring approach focuses on detecting changes in individual or interconnected process parameters compared to previous conditions. This allows the system to flag abnormal behaviour in the moulding process and support timely interventions minimizing disruptions and extending tool lifespan. 

WHAT IS EXPECTED? 

​

The experiment envisions a smart and responsive monitoring system that not only detects faults but also learns from operational feedback. By integrating human-in-the-loop capabilities, the system will incorporate input from operators to refine the fault detection logic and improve classification accuracy over time. 

Additionally, curated data will be organized in a structured environment, enabling easier sharing with R&D teams for further development. With AI-based models adapted to real production conditions, POLYCOM aims to strengthen process stability, support early fault diagnosis, and lay the groundwork for more intelligent, self-improving manufacturing systems.