Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB 
Project background
Trützschler Card Clothing GmbH is the leading manufacturer of high-performance card clothing for carding machines and roller carding machines used in the international textile machinery industry. The process is extremely complex and requires cutting precision of less than 0.5 mm at speeds of several meters per second.
Project description
The “Machine Learning for the Textile Industry” (ML4Tex) project optimizes this process by predicting the quality of the final product during production. It also focuses on the entire manufacturing cycle and uses a data-driven approach to identify resource-efficient, environmentally friendly solutions.
Data from sensors and raw materials along the production line is combined with quality data from Trützschler’s quality assessment department. Machine learning models convert production parameters into quality results, and the insights gained are displayed on a dashboard in the production area to enable immediate adjustments.
In addition, a high-resolution camera system with a resolution of <5 µm is being developed. This camera will be capable of detecting a wide range of wire defects online, something that is not yet possible today.
Current project status
The technical infrastructure is already in place, key data is being collected, and the first implementation phase has begun to deliver measurable added value to Trützschler, which is generating interest in the next phases.