Taking (quality) control

On July 18, 2017, the Washington-based Institute of Scrap Recycling Industries (ISRI) said China had notified the World Trade Organization of its intent to ban the import of mixed paper, “most scrap plastics” and metal slags and drosses by year-end. ISRI said the plastic scrap grades covered by the proposed ban include “polymers of ethylene, styrene, vinyl chloride and PET [polyethylene terephthalate].”

ISRI President Robin Wiener said a ban on imports would be “catastrophic” and have a “devastating impact” on the global recycling industry. While she said the import ban “appears to be more of a market issue rather than anything else”—with China’s government trying to boost its own scrap collection infrastructure—that country’s current National Sword campaign is concerned with quality and smuggling.

China’s National Sword campaign to crack down on contaminants in incoming loads of recyclables launched in early February 2017. This followed China’s Operation Green Fence, which ran through most of 2013, which put pressure on recyclers outside of China to boost the quality of materials exported to that country.

Most recyclers in the U.S. agree that these campaigns have helped to clean up quality in baled recyclables produced at material recovery facilities (MRFs). Yet, a complete ban on specific commodities is startling. China has made it clear that it will not continue accepting contaminated loads; recyclers must clean up their acts.

As U.S. recyclers continue to work toward producing cleaner bales more efficiently, the advent of robotics and artificial intelligence in sorting applications is gaining interest. Some say robotics offer reliable, consistent quality on the back end while also providing additional data and sorting multiple material types effectively.

Vision systems see clearly

Quality has been a concern for many MRFs since they first started up their conveyor belts. Tightening regulations affect recyclers every day. In light of contaminated streams from residential curbside recycling programs, MRFs daily sort through many nonrecyclables, from dirty diapers to garden hoses, in their efforts to sort and bale materials of value.

Workers in quality control stations, together with sorting equipment that includes optical sorters, ballistic separators, screens and eddy current separators, sort commingled materials.

One highlight of robotic systems is their ability to recognize and sort multiple types of materials—at one time.

“Robots are actually quite good at seeing the material differences,” says Timo Taalas, CEO of ZenRobotics Ltd., Helsinki.

He adds, “With robotic sorting, you can create very pure material streams.”

Matanya Horowitz, founder of Denver-based AMP Robotics, says there are two pieces to the company’s robotic system: the robot itself and a vision system, which is installed over a conveyor belt and is capable of understanding and learning all materials that pass through a MRF. Once material is recognized by the vision system, the robot uses suction cups at the end of a gripper to grab on to the item before releasing it into a bunker.

Horowitz says AMP’s robots can do more than differentiate between plastics Nos. 1 and 2, as one example. “The kind of vision system we’ve created can be more advanced—just like a person can tell the difference between a Coke bottle and a Pepsi bottle, it can too,” Horowitz explains. “It can identify and separate out very specific materials. A more pertinent example is separating out the same plastic resin but [in] different colors.”

The rules related to quality have changed, and if the new reality calls for strict quality requirements, then “the market is unprepared,” says Rich Reardon, managing director of Max-AI for Bulk Handling Systems (BHS), Eugene, Oregon. Reardon describes Max-AI technology as a “material identification platform” that is so far made up of two components, much like AMP Robotics’ technology: the Max-AI Vision System, which provides information, and the Max-AI Autonomous Quality Control (AQC) unit, which adds a robotic sorter to pick up as many as six different materials from the conveyor belt and place them in designated bunkers.

The first commercial AQC unit is already in operation at Athens Services’ MRF in Sun Valley, California. Integrating with the company’s existing NRT optical sorters, Max provides a fully autonomous PET sorting solution, says BHS.

Reardon says the unpredictable nature of the material stream is and will remain the biggest challenge for MRFs to overcome. Investing in these emerging technologies will prove to be worthwhile.

“MRFs will be able to run longer with lower operational expenses, produce more products with increased purity, capture accurate data for reporting and dynamic optimization and adapt over time to the changing material mix without major capital expenses,” Reardon says.

Every manufacturer interviewed for this story says its technology features machine learning capabilities, where the artificial intelligence collects information in real time as materials move along a conveyor belt. This trainable part collects data that recyclers have not had access to before, they say.

Reardon echoes Matanya’s point about the recent advancements of this technology and its sorting capabilities. The flexibility of this equipment, Reardon says, makes it possible to create new grades.

In addition, this technology provides flexibility in limited spaces, Taalas says.

As an example, Max-AI is no longer constrained to sorting PET versus high-density polyethylene (HDPE), Reardon says, “but can pick out PET trays from PET bottles while also grabbing aluminum, residue and mixed paper in the same location. Instead of being constrained by color analysis, Max can be trained to leave brown paper while removing cardboard.”

He continues, “Max can pick six items in one location. This would be extremely difficult for a person to keep track of with limited reach and time to make decisions.”

Structural changes

Robotic systems with artificial intelligence can multitask where humans cannot. When it comes to the topic of replacing workers with robotic sorters to improve consistency, recyclers and equipment suppliers are on both sides of the fence.

Horowitz sees potential in helping to staff positions in MRFs that are typically shorthanded. By incorporating structural changes within a MRF, “facilities can get more value out of people and include robots,” he says.

Horowitz adds, “When I talk to facility owners and operators, they say, ‘This won’t replace a person, but I will have a baseline capacity that I can count on even with all of the turnover I have.’”

Robotics also can help MRFs do more with the same number of employees. This was the case for AMP Robotics’ first installation. The Carton Council of North America is conducting a pilot program that uses artificial intelligence to improve the efficiency and effectiveness of carton recycling at Denver-based Alpine Waste & Recycling. (Read the cover story, “Forward focus,” beginning on page 32.) AMP Robotics and Alpine have used a robotic system, the AMP Cortex, nicknamed “Clarke,” to identify food and beverage cartons and separate them from the recycling stream.

Brent Hildebrand, vice president of recycling at Alpine, says the company moved the worker who was sorting cartons to a different area of its MRF, and now two material streams are cleaner than they were prior to Clarke’s installation.

Yet, Hildebrand recognizes the benefits of using robotics in MRFs in regard to workers. “As the labor market does get harder and harder to find the right employees in the MRF environment, it is a nice technology that’s reliably sorting, and it’s nice to be able to fall back on it.”

Kerry Getter, CEO of Austin, Texas-based Balcones Resources, says he is very much interested in robotics sorting. His intentions are to reduce operating costs and address staffing issues. With two shifts at its MRF, Balcones always has had problems staffing its second shift, Getter says. Robotics could help to alleviate this issue.

Getter explains, “If you run two shifts, the robot doesn’t care if it’s working a second shift, and we run two shifts at our facility. Staffing for a second shift has always been problematic; we don’t use temps in our company for a variety of reasons, so it was important to us in that regard.”

MRFs have fixed costs “and labor costs have gone up and will continue to do so,” Getter says. He adds that Balcones is working to mechanize as much as possible.

A fully automated MRF is the future in BHS’ eyes, Reardon says. Moving forward, he says BHS is challenging itself to integrate Max beyond quality control to create a completely automated MRF.

“We’ve already integrated Max into our controls, but eventually Max will control the entire process, a vision we call the Max-AI Autonomous MRF,” Reardon says.

In the Industry Leaders Q&A, beginning on page 50, James C. Fish Jr., president and CEO of Houston-based Waste Management, says he continues to press equipment manufacturers to come up with improved recycling equipment, such as robotics. While automation “gives us a higher yield on recycling,” he says he does not “expect that you’ll see labor being eliminated from the recycling business.”

Wait and see

Getter says he and other recyclers he has talked to like robotic sortation. However, they want to ensure it is proven before investing. Balcones, like most operators, performs return on investment (ROI) analyses on every piece of equipment considered for installation, and the technology is still so new that concrete information is not available from many manufacturers.

“As soon as the performance can meet the ROI criteria, we’re all over it,” Getter says. “I hope we’re buying machines in another year or so.”

Horowitz says he has been encouraged by how supportive the recycling industry has been of robotic sortation.

When the Closed Loop Fund (CLF), an investment fund that finances recycling infrastructure and sustainable manufacturing technologies to advance the circular economy, announced in March 2017 its request for proposals (RFPs) for projects that collect, sort and/or process postconsumer polypropylene (PP) plastic, it cited robotics as one field it would consider supporting.

Horowitz says, “The Closed Loop Fund has expressed a great deal of interest in helping facilities to get our equipment, so this is another place where the industry itself is helping us in ways I never would have dreamed of.”

Reardon says BHS plans to prove its artificial intelligence technology to some of the “wait-and-see owners and operators.”

Ultimately, robotic sortation with artificial intelligence could help to keep recyclers in the game, especially when it comes to bale quality. As Hildebrand says, “As more of the regulations on quality become a regular practice as far as what’s allowed and what’s not allowed in certain commodities, I think robotics will reliably sort the right materials. The quality specs will be much greater in the long run.”

The author is associate editor of Recycling Today and can be contacted by email at mworkman@gie.net.