USA - TOMRA Recycling, a leading global provider of sorting solutions, has announced the expansion of its GAINnext deep-learning-based artificial intelligence ecosystem with the introduction of two new applications specifically designed for material recovery facilities (MRFs) and secondary recyclers in North America. 

This latest enhancement focuses on improving the sorting capabilities for PET and paper materials, addressing common challenges faced by recyclers in the region.

The new applications—GAINnext PET Cleaner and GAINnext Paper Cleaning—utilize advanced deep learning technology to identify and classify hard-to-sort objects rapidly. 

These tools significantly reduce the reliance on manual sorting, which can be labor-intensive and inefficient. 

Capable of recognizing thousands of objects by both material type and shape within milliseconds, the GAINnext applications aim to minimize impurities in PET and paper bales, creating new opportunities for revenue generation, enhancing profitability, and lowering operational costs.

Integrating advanced technology for superior sorting accuracy

Central to the functionality of the new applications is an intelligent system that integrates multiple sensor data sources, offering greater sorting accuracy compared to traditional optical sorting methods. 

This integration complements TOMRA's AUTOSORT sensor-based material identification system, enhancing the high-purity performance of the GAINnext ecosystem. 

As a result, recyclers can expect maximized recovery and purity of valuable PET and paper materials, all while maintaining high throughput speeds.

Ty Rhoad, TOMRA Recycling’s Vice President of Sales for the Americas, highlighted the benefits of the GAINnext integration. 

“Recyclers can incorporate GAINnext into their existing lines, allowing them to enhance PET and paper recycling recovery without the need to install additional lines. This is particularly advantageous for operations with limited space,” he stated. 

The applications can process material at speeds of up to 2,000 ejections per minute, achieving throughput rates that can be up to 33 times more efficient than manual sorting.

Focused solutions for cleaner recycling streams

The GAINnext PET Cleaner application is engineered to sort opaque PET bottles effectively, enhancing the recycling process by efficiently removing challenging materials such as opaque white packaging and textiles. 

By utilizing deep learning technology, the system can eliminate over 92% of opaque objects treated with titanium dioxide, thus addressing common downstream recycling issues. 

It also improves sorting performance for transparent and colored PET materials by removing polyester textile waste, ultimately boosting the purity of recycled PET fractions.

On the paper side, the GAINnext Deinking/Paper Cleaning application is tailored for sorting office paper, newspapers, and magazines. Leveraging multi-sensor integration, this application effectively filters out impurities, including pizza boxes and egg cartons, ensuring a cleaner paper stream. 

It also distinguishes and removes greyboard at high throughput speeds, optimizing the quality of recycled paper products.

Indrajeed Prasad, Product Manager for deep learning at TOMRA Recycling, remarked on the application’s potential to improve cardboard sorting efficiency. 

“Our GAINnext Deinking/Paper Cleaning application enhances sorting performance for cardboard-based materials, creating high-quality revenue streams for paper recycling operations,” he noted.

Legacy of innovation in recycling technology

TOMRA introduced its first deep learning AI technology for recycling in 2019, initially targeting the identification and removal of polyethylene silicone cartridges from PE streams. 

Since then, the company’s deep learning engineers have trained artificial neural networks using millions of object images to tackle complex sorting challenges across various materials, including wood, plastics, and used beverage cans (UBCs).

In early 2024, TOMRA launched several innovative applications in the European market to efficiently separate food-grade from non-food-grade PET, polypropylene (PP), and high-density polyethylene (HDPE) at exceptional throughput rates. 

The addition of the PET Cleaner and Paper Cleaning applications in North America further expands the GAINnext ecosystem, ensuring that recyclers in the region can benefit from advanced sorting technology tailored to their specific needs.