(ORDO NEWS) — Scientists from Denmark have trained a neural network to distinguish between all major types of “household” and engineering plastics for sorting and processing.
The advent of plastics has brought revolutionary changes to industry and everyday life. These artificial polymers can have a wide variety of properties, which makes them ubiquitous.
But the more difficult it is to separate them from each other for recycling and reuse. The new method, proposed by scientists from the University of Aarhus in Denmark, makes it possible to recognize 12 different types of plastic on the fly. Mogens Hinge and his co-authors write about this in an article published in the journal Vibrational Spectroscopy.
As a rule, the difference in density is used to separate plastics: some materials quickly sink in water, others sink slowly, and others float altogether on the surface. But such a simple approach does not always work, and then you have to resort, for example, to near-infrared spectroscopy, which allows you to determine some types of plastics when material moves along a conveyor belt. The new method works with more polymers and with higher precision.
For recognition, data from hyperspectral imaging in the infrared range is used. They are analyzed by artificial intelligence trained to identify the signal corresponding to key types of plastics: polyethylene, polypropylene, polyethylene terephthalate (PET), polyvinyl chloride (PVC), polystyrene, and so on – a total of 12 options.
“Thanks to this technology, we can see the differences between all common types of plastic for consumer products and some high-performance types. We can even notice the difference between polymers, which are made up of the same building blocks, but form a different structure, ” said Professor Hinj.
The high recognition accuracy demonstrated by the system makes it possible to bring the purity of the material up to 96 percent, which is required for sending for processing. According to Danish scientists, the technology has already passed successful pilot tests. And starting this spring, it will begin testing it at local waste management facilities owned by PLASTIX and Dansk Affaldsminimering Aps.
Contact us: [email protected]