RETRO-Q

RETRO-Q ("Resilient Retrofit Sensing and LLM-Supported Quality and Condition Monitoring for Legacy Machines") addresses the insufficient digital integration of long-lived machine tools, a problem particularly widespread among SMEs. The resulting lack of transparency leads to reactive maintenance, quality losses, and unplanned downtime. RETRO-Q closes this gap with a modular, manufacturer-independent sensor and AI system that upgrades legacy machines in a resource-efficient manner rather than replacing them, thereby supporting Re-X strategies in production.

The system captures acoustic signals, vibrations, as well as fluid and temperature data, with acoustic and vibration-based features serving as early indicators of wear, operator error, and unstable processes. All processing takes place locally on an edge device to ensure data protection, data sovereignty, and responsiveness even under limited connectivity. On this basis, AI models for anomaly detection, condition diagnosis, and resilient operation are trained, generating recommendations for process stabilization and rescheduling.

A key distinguishing feature is the integration of large language models as a dynamic, context-sensitive maintenance and operating manual: technical knowledge, experiential data, and historical operating data are made accessible through natural language, while operator feedback continuously expands the system's knowledge base and fosters a collective body of knowledge. The modular approach, built on configurable function blocks and open interfaces (OPC UA, MQTT), ensures adaptability to heterogeneous machine types and control generations.

The project (duration 01.01.2027–31.12.2028) is structured into sensor and software development followed by validation under real production conditions at partner companies. Starting from TRL 4, it targets TRL 6–7. RETRO-Q strengthens the resilience and competitiveness of industry in Rhineland-Palatinate, lowers investment barriers, and — through a planned spin-off — establishes a regional competence hub for resilient digitalization solutions in mechanical engineering.