AI-Powered Plastic Recycling with Artificial Neural Twin
In the K3I-Cycling project, 16 partners are developing methods to recover plastic recyclates with defined quality levels from mixed lightweight packaging waste
Fraunhofer LBF is responsible for the work package on recycling and recycled material production and focuses on laboratory and pilot-scale tests for recycled material recovery, as well as the development of additive packages, including bio-based stabilizers. Using materials analysis and machine learning, polyolefin recyclates are classified according to aging and contamination, and quality clusters are formed, which are incorporated into standardization activities and digital product passports.
In the AI application hub, contaminants are detected and sorted out at the item level. Analytical data from Fraunhofer LBF supports the evaluation of sorting progress and the further development of DIN SPEC 91446. The Artificial Neural Twin identifies opportunities for AI-supported measures such as battery detection or optimized logistics. Additives such as odor absorbers and compatibilizers are intended to improve the processability and long-term stability of the recycled materials and enable their use in demanding applications, such as food-contact packaging films.
