Michigan MRF benefits from the addition of robotics, AI

The material recovery facility (MRF) owned by Resource Recovery and Recycling Authority of Southwest Oakland County (RRRASOC) in Detroit and operated by Phoenix-based Republic Services added artificial intelligence (AI) and two robots to its processing system this year that have helped to improve recovery rates and sorting capacity in addition to providing workers with valuable skills in robotics and AI.

The 50,000-square-foot MRF, which was destroyed by a fire in 2014, was rebuilt and began processing material in 2016, says Mike Csapo, RRRASOC general manager. It receives 200 to 300 tons per day, Monday through Friday, and commonly runs one shift on the weekend. This year, RRRASOC added two robotic sorters and AI supplied by California-based Glacier to its existing processing system, which includes three fiber screens, two optical sorters, an eddy current and a magnet. 

“The robotics are all brand new as of this year,” Csapo says, adding that the first robot was installed in January and the second in July. 

RRRASOC learned about Glacier when the robotics company participated in Nextcycle Michigan’s inaugural cohort in the Recycling, Innovation and Technology Track.

Rebecca Hu, CEO of Glacier, says the company was introduced to many of the MRFs and waste agencies across the state through the program.

“We found this opportunity to partner with RRRASOC and SOCRRA [the regional recycling authority for the cities of Berkley, Beverly Hills, Birmingham, Clawson and Ferndale] … because they were both facing operational challenges like labor shortages,” she says.

Glacier partnered with the RRRASOC on a grant proposal through Michigan’s Department of Environment, Great Lakes, and Energy (EGLE).

“We ended up getting funded by not only EGLE for the deployment of multiple Glacier robotics sorting systems and AI data sets, but we also received additional participation from other third-party grant funders like The Recycling Partnership, Carton Council and the Foodservice Packaging Institute,” Hu says.

RRRASOC has deployed its Glacier robots on its container line to help pick polypropylene (PP) and on its residue line to recover recyclables that otherwise would have been lost to residue.

“That wasn't always happening because of labor challenges getting enough people on the sort line,” Csapo says of sorting PP. “To be able to deploy a robot to do a job for which we couldn't get a person is really important.”

On the second-chance line, Hu says, everything the robot does not pick would likely end up being landfilled.

“That robot is a great showcase of the flexibility of the system because it's picking PET [polyethylene terephthalate] bottles, HDPE [high-density polyethylene] color and natural, aseptic cartons, gable-top cartons and aluminum cans, so there's a huge wealth of items that we can go after this robot,” she says.

Through the deployment of its AI, Glacier found that the most common recyclable on RRRASOC’s residue line was PET bottles.

“Based on the amount of PET that was available to be recovered, we were actually able to quantify the incremental commodity revenue opportunity,” Hu says. “And, from there, RRRASOC was able to make a couple of staffing changes upstream because they realized that the labor cost to go and recover that PET was well worth it in order to divert more of those tons from landfill.”

Hu adds that the robot on the MRF’s residue line is recovering nearly 600,000 recyclables, or 20,000 pounds, every month that would have gone to landfill but now is making it to the market, enabling a payback period within the first year.

“I often tell our MRF customers that if you want to start trying out AI data in your facility, the best place to start is on that residue line because it gives you a real-time pulse on how your facility is doing in areas for further recovery,” Hu says. 

Glacier’s technology appealed to RRRASOC in part because its robots are purpose-built, smaller and modular, Csapo adds, noting they are more directly tailored to specific sites and for specific use cases. 

“One of the things that we liked about Glacier’s installation process is it was able to be done in a way that didn't interfere with production and basically done over the course of the weekend,” he says.

“We also were intrigued by … their ethos of wanting to bring AI and technology to the materials management space in a way that hasn’t been done before.”

Hu says the installation is helping the MRF with the physical robotic recovery of additional items and also providing “a real-time feedback loop and visibility into what's coming into the facility and additional opportunity to improve the facility’s operations and yield.”

Workforce development is another beneficial aspect of the installation, Csapo and Hu say. 

“One pillar of Glacier’s approach is that successful integration of our technology into MRFs also requires that the folks on the ground are empowered to know how to use our technology,” Hu says. “We believe it is incumbent upon us to teach our customers and their employees how to best use our technology. So, whenever we install our robots at a MRF customer site, it always comes accompanied by a robust series of training sessions that our team will do with the on-site maintenance and repair crew to make sure that they're well-equipped to troubleshoot, to maintain and to figure out our robots themselves. And in doing so, we also equip them with important skills on how to interact with this type of technology and automation in the future because we believe over time that you're only going to see more of this type of automation entering these facilities.”

“I'm seeing some remarkable collaboration between Glacier and our operators,” Csapo adds of the Republic Services staff at the MRF, noting strong working relationships have developed between the MRF’s maintenance and supervisory teams and Glacier’s support staff.

“None of the folks in our plant had been exposed to robotics or AI or any of this technology, and we're seeing them really learn what it means to install, operate, manage [and] repair this type of equipment and to the extent that this type of technology is going to be increasingly important in the future, it's really valuable to have a team that has hands-on experience and understands the ins and outs of this technology without relying solely on the outside support of the vendor’s tech team.”