A camera system combined with artificial intelligence that can automatically identify defects in the production of carbon fibres has been developed at the RWTH Aachen Institute of Textile Technology (ITA) in Germany.

The development, which promises to make the expensive manual inspection of carbon fibres obsolete and reduce their production price in the long term, earned young engineer Deniz Sinan Yesilyurt the second prize in the VDMA’s 2022 Digitalisation in Mechanical Engineering Young Talent Award. The annual prize is awarded for outstanding theses from institutes in Germany, Austria and Switzerland.

Carbon fibres are very light, weighing up to 50% less than aluminium. Their combination of low weight and good mechanical properties offers many advantages and in times of energy transition, the lightweight materials that are created with them are more relevant than ever before. At the same time, carbon fibres can be as resistant to external stresses as metals, but achieving these valuable properties is very complex.

Up to 300 individual fibre strands – bundles of individual fibres – have to be monitored simultaneously during production. If carbon fibres tear, it costs time and money to sort out the damaged fibres. Many other defects can occur in the fibres during production.

Yesilyurt attached a camera to the carbon fibre line that takes pictures of various fibre defects during production and collects them in a database. The artificial intelligence in the camera’s information technology system evaluates the defects by assigning the images to predefined reference defects. In doing so, it recognises various fibre defects with a classification accuracy of 99%. The process can also be used with other chemical fibres.

www.ita.rwth-aachen.de