In a production environment with an output of over 55 million clips per year, there are considerable challenges in quality assurance: Manual inspection is only carried out sporadically between measurements and troubleshooting on the packaging line. As a result, visual inspection is often delayed or not carried out at all.
The result: bottlenecks in the production flow, which lead to an intermediate buffer in the automated small parts warehouse (AKL) and cause additional trips.
Without timely inspection, there is a risk of major damage to the line and longer downtimes. Automated process monitoring is necessary to counteract this and relieve the inspection staff.
The presented AI solution uses an integrated camera system with a direct connection to the machine control system. AI-based error analysis detects deviations and anomalies inline using specialized deep learning models.
The real-time evaluation and visualization ensures an automated IO/NIO product split and provides precise operating instructions via an individually developed browser-based software solution.
Key technical features:
Automated testing significantly relieves testing personnel and ensures smoother servus runs. Maximum transparency is achieved through graphical evaluations of AI results including date, time, WT/ARC number, transport source and fault classification.
The solution improves the performance of the entire system, increases the overall efficiency of R-Clip automation and lays the basis for further AI-based optical testing in other production areas.