In the coming years, the amount of batteries that will have reached the end of their service lives is set to increase exponentially. As batteries are made from critical resources, disassembly and recycling are of great importance. In order to make these scalable, automation is required. The Swiss Battery Technology Center (SBTC) is taking up the challenge.
Picture: Innosuisse
Batteries are fundamentally important for the digital economy, for greener mobility and for the continued development of renewable energies. However, battery manufacturing has a serious impact on the environment. Battery recycling could make a valuable contribution to improving the environmental footprint of batteries, while also helping to bridge the gap between demand for resources and actual supply.
A typical lithium-ion battery for powering a mid-range electric vehicle weighs around 400 kilograms. Just under a third of such a battery is aluminium, a good 15 percent graphite, ten percent electrolytes, ten percent nickel, five percent copper and five percent plastics. The rare metals lithium and cobalt each make up around two percent of the total weight. The resources in batteries are vital to the energy and transport transition. Europe only has very limited quantities of such resources itself though, so the continent is dependent on imports. The EU categorises the majority of resources needed for battery manufacturing as critical, because there is no security of supply. For this reason alone, there is strategic interest in keeping these resources in circulation at the end of a battery’s service life. This also means that spent batteries should be turned into new ones wherever possible. However, the disassembly this requires poses a number of challenges.
The only way to reduce extraction is to maximise recovery of individual resources. Moreover, resources can only be reused in battery production if they exhibit a high post-recycling quality standard. The recycling and reprocessing procedure should be designed to be as energy-efficient as possible: Avoiding thermal processes, such as the melting-down of materials, ensures good ecological compliance.
The number of vehicle batteries that will have reached the end of their service lives, possibly after secondary use as stationary storage units, is set to increase sharply, or even exponentially, in the next 20 years. By 2035, Europe will already require recycling capacity for around 900,000 tonnes of lithium ion batteries from vehicles and other applications. It is therefore necessary to develop scalable procedures.
The Swiss Battery Technology Center (SBTC) at Switzerland Innovation Park Biel/Bienne is tackling these challenges in collaboration with two other organisations. The Eurostars project LAMBDA aims to develop a system that will train robots to disassemble electric vehicle batteries with the aid of state-of-the-art AI. The aim is to minimise the use of processes that destroy the batteries. This is a challenge in terms of both automation and robotics: On one hand, the robot must learn to handle batteries from a very wide variety of manufacturers. On the other hand, the batteries and the components they contain have completely different form factors, vary in terms of rigidity, and are often bonded, welded or moulded. After ten or fifteen years of use, they sometimes show considerable signs of wear and tear, and may be corroded or soiled. Disassembly of batteries also entails considerable risks. If they are handled incorrectly, they may release toxic fumes or cause fires. “Thanks to AI-assisted robotics,” says Christian Ochsenbein, Head of the SBTC, “people are involved in fewer hazardous activities and, at the same time, the important resources are recovered in an energy-efficient manner.”
Today’s robotics cannot cope with the complexity of the task. However, machine-learning procedures bring a promise of improvement. In this project, the task is first simulated in a virtual, photorealistic and physically correct environment, and the robot is trained to perform the necessary steps using highly detailed digital battery models. This includes independent recognition of components, navigation, and manipulation of the individual parts. For safety reasons, the tasks learnt are then tested on dummy batteries in a real environment. This allows the robot’s capabilities to be honed before trials involving automated disassembly of life-expired conventional vehicle batteries, which the SBTC wants to start in the near future.
