Extrusion International 3-2026

26 Extrusion International 3/2026 MEASUREMENT TECHNOLOGY Introducing New AI-Driven SIMAC® Surface Testing System Zumbach Electronic is advancing surface inspection technology with the launch of the next-generation SIMAC ® Surface Testing System, now empowered by AI-driven Intelligence. Designed for cables, tubes and hoses, the new SIMAC® solution delivers a breakthrough in accuracy, adaptability and on-line quality assurance. Built on deep-learning innovation, the system moves beyond traditional defect detection to provide manufacturers with actionable insights, continuous learning capabilities and an unprecedented level of process transparency. On-line detection: Identifying defects at full production speed At the core of the new SIMAC ® platform is its arti cial intelligence-based high-performance detection engine, developed to recognise surface abnormalities instantly and reliably, with the following bene ts: • Detects surface defects inline at full production speed • Requires no defect labelling, because it is trained entirely on defect-free images • Provides adaptive behaviour, automatically compensating for process drift • Allows fast model recalibration for new production conditions • Includes adjustable sensitivity, enabling a con gurable precision/recall balance depending on customer requirements This means that operators no longer need to manu- ally tune thresholds or de ne defect rules. Instead, the deep-learning engine continuously evaluates the product surface, agging any deviations from expect - ed patterns. Intelligent classi cation: Knowing exactly what is wrong Detection is only the rst step. With SIMAC®’s new intelligent classi cation module, users also gain clarity about what type of defect has occurred. The system can: • Automatically assign defects to prede ned types • Provide precise defect localisation on the product • Support the introduction of new defect types, when combined with model retraining • Offer adjustable sensitivity for enhanced precision or broader coverage • Apply a “one-for-all” model, meaning a single trained model can serve multiple products This unique AI-driven defect classi cation tells us ex - actly what is wrong, enabling engineers to better un- derstand production issues rather than just reacting to alarms. Continuous improvement: A system that evolves with production One of the most forward-looking capabilities of the new SIMAC ® is its ability to evolve. The system’s contin- uous improvement framework allows it to become in- creasingly accurate, consistent and adapted over time. Key features include: • Fast recalibration of the detection model • Re-classi cation and correction of mis-detected defects • Incremental enrichment of training datasets • Full retraining when introducing new defect categories • Ongoing optimisation of thresholds and system behaviour Because the new SIMAC ® is not based on xed rules, it evolves with production, positioning itself as a fu- ture-proof solution that grows alongside customer pro- cesses, raw materials and quality requirements.