In the latest issue of ZENTEC’s Zukunftsmagazin (Future Magazine), CORDENKA shows how AI models developed in-house increase transparency and efficiency in our production processes – from visualizing plant status to automated fault detection.
A great example of how practical knowledge and modern technology go hand in hand! 🤝🏼
You can read CORDENKA’s article below:
Artificial intelligence that takes the strain
Use of image recognition in production at CORDENKA
CORDENKA GmbH & Co. KG, based at the Obernburg Industrial Center, successfully uses AI models developed in-house to visualize the status of its systems. The model, an idea from the workforce, was developed as part of the “AI Transfer Plus” funding project of the Bavarian State Ministry of Digital Affairs in collaboration with the Technical University of Aschaffenburg and SPIE Automation in Niedernberg. The aim was to increase the transparency and availability of relevant process information in order to be able to react to deviations at an early stage and design operating processes efficiently.
The close integration of practical knowledge from production, state-of-the-art AI methodology, and practical implementation in the plant has resulted in a solution that is not only technologically advanced but also brings directly noticeable relief to the everyday work of employees.
Initial situation
CORDENKA is the world’s leading manufacturer of industrially biodegradable rayon yarns. In textile post-treatment, spools are washed and dried. To do this, they are placed in a drain rack and knotted together with the previous spool to form a continuous thread. With over 2,000 discharge points, this is quite challenging and stressful for employees to place the next spool in time to avoid efficiency losses. In the past, work was organized by visually checking the condition of the gate by the employee, who then planned the optimal time window for the activity.
Central shift planning
The use of artificial intelligence has significantly simplified the organization. With the support of SPIE Automation and the Technical University of Aschaffenburg, employees were able to train AI models to recognize the fill level of the spool and make this information available to employees via monitors and on-site LED displays. The data generated is collected in a database. With this information, it is only a small step to centrally plan the shift scheduling of employees. A significant relief for supervisors and employees!
Continuous image and status data
The second challenge, besides software development, was practical implementation. The solution, developed under project manager Roland Johannes, was surveillance cars suspended on tracks, each equipped with four webcams to analyze the spools of the gates—an overall cost-effective solution with enormous added value for CORDENKA. The monitoring device provides continuous image and status data that is processed in real time, enabling bottlenecks or anomalies to be detected and documented at an early stage. Integration into existing alarm and maintenance systems enables proactive maintenance and more reliable production planning.
Fast troubleshooting, better traceability
The experience gained from the funding project enabled further issues to be addressed. Thread breaks occur during the process, which are detected in a thread cluster of over 700 individual threads via a light barrier as a collective message. The employee’s task is to locate the broken thread and replace it. This is a task that requires a great deal of manual skill, experience, and practice. The use of webcams and a new AI model has automated the detection of the position of the broken thread. Visualization using programmable LED strips in the system guides employees directly to the fault so that it can be quickly rectified. The data collected helps to identify and reduce areas where faults are most likely to occur. The new system significantly reduces the risk of tangles caused by broken threads being caught by the running machine. This not only increases productivity, but also improves occupational safety and the working atmosphere, as recurring manual search and recovery processes are minimized. In addition, AI-supported visualization and diagnostic solutions enable better traceability of production quality.
Employees benefit from a clear overview of the plant status. This reduces training requirements, as new colleagues can quickly understand and operate the systems.
Further steps to follow
With the first AI solutions implemented in production, new opportunities are becoming apparent and the newly established AI group is receiving inquiries from various areas of the company. A real asset!
