AI-based material analysis for the circular economy

The recovery and recycling of valuable materials from a wide variety of sources is becoming increasingly important, both to help meet global climate targets and to respond quickly and cost-effectively to customer requests in manufacturing plants. Particularly when using recycled raw materials, it is important to have precise knowledge of the quality of the recyclate, i.e. the composition of the material flow.

JOANNEUM RESEARCH DIGITAL has established a focus on material characterisation in the recycling sector and has laid the necessary foundations in the hyperspectral laboratory. Sensor systems that go far beyond the range of human vision enable contactless detection and subsequent analysis of recyclable materials. Characterisation takes place at all stages of the recycling chain, e.g. when sorting residual materials or identifying contaminants, but also as an incoming inspection to ensure the necessary raw material quality for production, thus closing the loop.

The current project portfolio covers different material fractions and positions in the cycle:

• InSpecScrap (Zukunftsfonds Steiermark) develops a characterisation of foreign substances in steel scrap as an input control for steel production
• StraTex (FFG-Call Circular Economy) analyses textile fibres with regard to their composition.
• circPLAST-mr (FFG-Call ENERGY OF THE FUTURE, circular economy) - Lead project on the topic of plastics recycling
• KRAISBAU - Lead project on the topic of construction waste
• KI4ResOpt serves as the basis for research and development of the necessary sensor technology, including customised lighting and sensor fusion, as well as AI-based analytics using modern deep learning architectures and explainable AI components.