TECHNOLOGY | SORTATION
facility, also owned by RDS, currently processes 150 tonnes/day of local MSW with over 90% uptime. AMP says the operational efficiency is an industry first, delivering an “unprecedented level of reliability for mixed waste sorting systems at a scale and footprint not previously feasible economically.” The AMP One system separates bagged trash into its component parts of mixed recyclables, organics, and residue. The RDS facility can divert more than 60% of landfill-bound material with the system, which can effectively sort dirtier materials with more contaminants.
Above: AMP has developed AMP One, a highly auto- mated
factory-scale sortation solution
automated factory-scale sortation solution, avail- able in a modular design capable of achieving throughput of 15,000-100,000 tonnes/year with 90%+ recovery of recyclables. AMP One uses AI-driven sortation technology to transform single stream municipal solid waste, materials recovery facility residue, mixed plastics and other infeeds into bales of valuable recyclable materials. A primary benefit of an AMP One system is
realised when a facility is built with AI capabilities in mind from the outset, said Spelhaug. “One can optimise its design around the technology, which supports changes measured in orders-of-magni- tude compared to a retrofit. It is not just that one may operate with fewer sorters but also without manual sorting altogether,” said Spelhaug. With the consideration of AI-enabled data collection up front, one has the possibility to make the sorting lines wider and higher speed. AI can help replace entire classes of equipment that create maintenance and downtime events. “By reducing the cost of sorting, our technology is growing the market of recycled materials by making it profitable to recover material where it currently is not. We create the opportunity to sort new material combi- nations, unlocking higher revenue from the same infeed material,” Spelhaug said. This year, Recycling and Disposal Solutions of
Virgina (RDS) installed an AMP One advanced recycling system at the Pitt County Recycling Center in Greenville, North Carolina, to process approximately 10,000 tonnes of single-stream and commercial recycling. RDS installed the unit at the site, which it owns and operates, as part of a modernisation effort to process recyclables more efficiently at lower cost through efficiencies. Also, this year, AMP installed an AMP One system in a municipal solid waste (MSW) facility, which Spelhaug calls an industry first. The MSW
22 PLASTICS RECYCLING WORLD | September 2024
Automated operations California-based tech start-up Everest Labs has developed an AI, robotics and data platform purpose-built for recyclers, reclaimers and packag- ing manufacturers designed to automate sorting operations, improve feedstock supply, and increase recovered recycled content. Its RecycleOS platform, which the company considers the first enterprise AI and robotics system for recycling plants, combines several proprietary company innovations in vision system technology, industrial edge computing, and deep learning AI to efficiently sort and process high volumes of recycla- ble materials accurately and quickly. It increases material recovery rates by up to 40% while lowering labour and material disposal costs by 40-60%, according to JD Ambati, founder and CEO. The RecycleOS platform can conduct live audits and identification of the recycling stream, provide object tracking for sorting equipment and is able to perform packaging type, size, and mass characteri- sation of an object through its enterprise level cloud data platform. The company built its material characterisation model and its machine learning model from the ground up. Everest Labs typically operates the RecycleOS
platform with robotic sorting equipment, that, when combined, outperforms manual sorting of bottles, containers, and other recyclables from commingled input materials by a factor of two or three, said Ambati. The company’s RecycleOS Robotics cell features a self-contained industrial 3D vision system that is self-lit and easily mounts on top of a conveyor to capture an image of the objects. The vision unit incorporates 3D depth sensing cameras that can identify up to two hundred items in each frame. The system can accurately process 30 frames/sec. Image and sensor data is sent to an industrial edge computer containing a custom AI model for object detection, developed by Everest Labs. The AI algorithm compares the images to a classifica-
www.plasticsrecyclingworld.com
IMAGE: AMP
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44
