Robot spinout to boost metal recycling
A spinout from the University of Liège in Belgium is tapping into a decade of research to boost the efficiency of metal recycling.
Matvision is combining robotic systems with X-ray, infrared, laser and 3D sensors and AI for single pass, accurate sorting of different categories of metals.
The research from the GeMMe laboratory of ULiège was developed with industrial partners COMET Group in Mons and Cilyx in Liège as part of the PICKIT project and the company is backed by the imec.start fund. The company is looking to raise €1m later this year.
The focus for Matvision is the dismantling of cars and household appliances, for strategic metals such as copper, aluminium and nickel. There are currently more than 350 car recyclers in Europe, and given the industrial developments based on recycling, the potential market for is even larger.
- Mecaware raises €40m for recucling battery scrap
- Deal for rare earth recycling in the UK
- €100bn for European cleantech, battery recycling
Matvision is also a service company that studies the feasibility of the most suitable sorting technology a case-by-case basis using its multi-sensor and multi-output sssytem. Together with its subcontractors, it then supervises the in situ assembly of the robotic sorting lines. Feasibility studies are carried out on the prototype line installed at ULiège.
“After ten years of development and validation with industrial partners, our technology is now mature. It stands out from other systems in the world because of our ability to process a very large quantity of various metal waste with great precision in a single pass. With Matvision, our industrial partners can shorten recycling loops and increase the value of the waste they process and resell, making our machines profitable in record time,” said Robert Baudinet, co-founder and CEO of the company.
Industrial demonstrator
The COMET Group carried out the first industrial validation of this robotic sorting system in 2021 with a €10m industrial demonstrator. MULTIPICK was a chain of 16 sorting robots with a sorting capacity of 20, 000 tonnes of metal waste per year (i.e. one billion parts at a rate of 16 parts sorted per second).
The technology is based on the combination of the development of sensors for recognising materials and high-speed robotic sorting, all driven by its own machine learning framework.
The X-ray measurements that allow real-time evaluation of material density while a 3D scanner measures the shape and volume parameters of crushed objects to choose the optimal gripping points for the robots with clamps.
A hyperspectral camera is used to measure the reflectance of materials in the visible and infrared spectrum while a LIBS (laser-induced breakdown spectroscopy) laser differentiates between different metal alloys.
After learning the different desired categories (aluminium, zinc, copper, brass, lead, stainless steel, etc.), each new fragment is assigned to a category. The conveyor belt rotates at more than 1 metre per second and the Matvision system can process up to 20 types of materials in a single pass. Current systems can only separate two families simultaneously.
“Only multi-sensor solutions such as that of Matvision will be able to sort increasingly complex products in the future. By positioning themselves as pioneers in robotic metal sorting techniques, Walloon engineers can hope to play a leading role in the development of the circular economy on an international scale,” said Éric Pirard, professor at the GeMMe (Géoressources, Génie Minéral et Métallurgie Extractive) lab.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
