Biogas

Biomasse

Erneuerbare allgemein

Energieeinsparung

Erdwärme

Grünes Geld

Klimaschutz

Kraft-Wärme-Kopplung

Meeresenergie

Mobilität

Photovoltaik

Solarthermie

Wärmepumpen

Wasserkraft

Windenergie

content-Partner

The Role of Multispectral Scrap Characterisation and Artificial Intelligence in Efficient Steel Recycling

Recycling of steel scrap is an important component in the modern steel industry. However, the composition of steel scrap is volatile and can contain a large range of unwanted particles, including non-ferrous metals, plastics, or concrete.

A good characterisation of the scrap composition is therefore key for an efficient steel recycling process. Hyperspectral imaging, covering a range far exceeding the visible light spectrum, delivers a spectral profile that allow to distinguish between ferrous and non-ferrous scrap particles. However, the sheer volume of information in hyperspectral images, makes it impossible for humans to inspect these images in an industrial process. Therefore, Machine Learning (ML) is used to segment the images into the regions covered by different materials. Deep Neural Networks (DNNs) are the machine learning algorithm of choice to do this, and the results show that the DNNs deliver a good segmentation performance. The results of the image segmentation are the basis for future application like optimised scrap compositions, reduced energy consumption, or even self-adapting steel recycling processes.