Recycling used batteries has long presented a logistical challenge for the industry because of the wide variety of shapes, sizes and chemical compositions involved.
Mixing incompatible battery types during the recycling process can significantly reduce efficiency and increase operational costs.
To address these difficulties, the battery recycling startup Redwood Materials has implemented a new automated sorting system powered by artificial intelligence.
The company was founded by Tesla cofounder JB Straubel to recover valuable materials such as lithium, nickel and cobalt to reduce reliance on mining and imports.
According to the firm, the newly developed system can identify batteries with more than 98% accuracy, as reported by “Business Insider”.
The process begins when old batteries arrive in a jumble of boxes and bins at the facility.
Machines then organize the items into a single-file line before they pass through a specialized tunnel.
This tunnel is equipped with various sensors and cameras designed to examine the label, weight and physical shape of each individual unit.
An artificial intelligence model, which was trained on the company’s own extensive battery data, then determines the specific type of battery.
Once identified, the system directs the battery to the appropriate location for the recycling process.
The entire operation runs on a cluster of graphics processing units located on site.
Redwood Materials reports that the system was constructed in only a few months.
The technology is currently capable of handling several tons of batteries every hour.
Industry analysts suggest this development serves as a practical example of how artificial intelligence can address messy, unglamorous industrial problems.
By streamlining these complex tasks, such automation could potentially make American manufacturing more competitive in the global market, according to “Business Insider”.
The company continues to focus on refining its recovery methods to ensure a more sustainable supply chain for essential battery components.
