When is it time to retrofit your MRF?

Maybe your recycling facility isn't producing the same quantity of saleable product as in the past. It could be that you need to increase the yield of a certain product that is in high demand. Or possibly recyclate quality is not where it needs to be to meet your customers' needs.

"From product quality to energy consumption to reducing maintenance, recyclers have a range of needs when considering a retrofit or technology upgrade," explains Gleb Trofimov, area sales manager for TOMRA Recycling. "And, as suppliers, the first thing we consider is the right technical solution to meet the customer's targets. It could be upgrading to the latest generation of optical sorter, incorporating a deep learning AI solution, or adding a sorter in the line. We consult with recyclers to make sure we offer the right solution to meet their goals."

Today, sorting automation technology available to material recovery facilities (MRFs) and plastics recovery facilities (PRFs) is rapidly evolving. New versions of optical sorters that are just 10 years old will have updated features that can offer more granular material sorting.

"If yields are dropping, landfilling costs are rising, or manual labour costs are increasing, the facility might benefit from updated technologies or add-ons that could help improve the sort," Trofimov adds.

Beyond improved sorting performance, looking at retrofits and upgrades to the line can help to lower long-term operating costs. Trofimov points to more recent improvements in newer machines that offer lower energy consumption and use less air. Plus, the reliability and easier maintenance designs help to improve efficiency at the facility. These combine to reduce long-term maintenance and operating costs for a better bottom line.

Data integration 

Optical sorters generate vast amounts of data that can be leveraged to improve overall operational efficiency. In the past, this information was tied to the individual machine, and a worker had to physically be on site to access the machine data for analysis. While this data served to help the one facility, sharing this data across multiple locations was challenging.

A recycler can significantly upgrade newer sorter performance by leveraging cloud-based reporting technologies currently offered by manufacturers. "Our secure cloud-based monitoring portal, TOMRA Insight, provides near-real-time sorting data as a strategic management tool that can be safely accessed anywhere an internet connection is available," says Trofimov. "And this is not just one machine. The data is offered by machine, by plant, and company-wide, so it provides reporting and monitoring for all TOMRA machines."

Sorters are now connected devices that transform sorting from an operational process into a strategic management tool. Customizable reports enable key personnel to analyze critical operational data, such as yield, throughput, and purity by sorting machine, by shift, by location, or for all facilities.

Not limited to production details, connected machines assist by reporting machine availability, performance, maintenance information, and active alarms. These alarms serve to keep recyclers ahead of potential issues and optimize sorter performance.

Additionally, newer deep learning artificial intelligence (AI) waste analysis tools like the PolyPerception Analyzer offer further transparency to operations with real-time material monitoring at key points in the sorting process. Every object passing under the camera is analyzed in real-time, and deep learning models quickly adapt to changes in the waste stream. Waste analysis delivers value to the facility through improving sorting efficiency, assessing true material quality for making commercial decisions, and certifying material quality for compliance.

AI add-ons

AI has grabbed much attention in the recycling industry over the past few years. While AI has been used for material separation with optical sorters for decades, the deep learning subset of AI is now expanding mechanical sorting's capabilities. The exciting trend is the pairing of these cameras and deep learning algorithms with sensor-based sorting capabilities to provide a granular sorting of materials.

"The basic understanding we need is that both are different sorting technologies, which have their own unique values," explains Indrajeet Prasad, product manager of deep learning for TOMRA Recycling. "Near infrared (NIR) and other sensors in optical sorters do the sorting based on material types, distinguishing between, for example, polyethylene terephthalate (PET), polyethylene (PE), and polypropylene (PP) in the case of plastics. Deep learning, on the other hand, sorts based on visual information — what the eye can see."

Combining the NIR sensors of units like AUTOSORT with GAINnext deep learning AI algorithms is the peanut butter cup of the recycling world that solves sorting tasks once thought difficult or impossible. It addresses several limitations of the camera-based AI solution alone when sorting key materials at the MRF or PRF.

For one, the pairing enables material to be sorted using valve-block technology much faster than robotic arms. "The arms are designed to mimic the picking action of a human, so they offer around 70 or 80 picks per minute," says Trofimov. "Valve blocks deliver a much higher capacity, and we are seeing purity rates in excess of 95 percent at throughput rates reaching six tons per hour."

Additionally, pairing sensor-based traditional AI with deep learning offers a more granular sort of plastics. "The camera with deep learning will see colour, shape, and texture of the material on the belt. So, it can recognize a green soda bottle or a clear water bottle, or a milk jug," adds Prasad. "When combined with the NIR sensor of the optical sorter, we can distinguish that it's a PET bottle with either a PET or PVC label."

Deep learning ushers in a new era of sorting capabilities and revenue streams for the recycler, including food-grade plastics sorting, upgrading PET or paper streams, advancing the sorting capabilities for aluminum, and providing a more sophisticated sort of wood materials.

However, Trofimov advises consulting with the technology provider to ensure the right combination of sensors are employed. "When retrofitting, it may be that upgrading to the latest optical sorter will help to reach established targets, or it might require an additional sorting machine. We are here to help guide recyclers through the considerations to ensure they get the combination right."

Retrofit consultants 

When it comes to machine retrofits, Trofimov explains that recyclers often have a triangular need. "Improvements typically focus on wanting to improve yield, purity, throughput, or a combination of all three." He also mentions that on the back end, recyclers are looking for easier maintenance and lower energy consumption.

Market changes can prompt recyclers to augment their facilities with newer technologies. Several years ago, thermoforms were not as plentiful in the PET stream as they are today. If a recycler wants to sell material for bottle-to-bottle recycling, PET thermoforms must be removed from the bottles.

"You cannot recycle a thermoform into a bottle, so you need the optical sorter to detect what is PET bottle and PET thermoform," explains Trofimov. "Just a couple of generations ago, our AUTOSORT units could not make this distinction, but sensor and algorithm updates in our latest generation AUTOSORT can separate PET bottles from PET thermoforms."

Trofimov recalls working with a plastics recycler in British Columbia, Canada, to replace the company's older optical sorters to assist in making its recycled plastic pellets. The upgrade to the latest sorters gives the owner real-time data to analyze the sort. Plus, the upgraded circuit incorporates deep learning AI to help further purify the PET stream.

"The owner values the information about the material that passes through the machine," he explains.

Additionally, a solid waste management company with operations across North America recently replaced its approximately 15-year-old optical sorters at its recovery facility in Ontario, Canada, to improve the recovery and purity of its old corrugated container (OCC) product. The motive, in addition to yield, was for easier machine serviceability.

So, when is it a good time for the MRF or PRF to consider an upgrade? Trofimov says that every facility collects its own statistics and measurements to make the determination. "It could be changing material streams, or the yield of a certain product starts to wane. Normally, it doesn't happen overnight. These decisions are made over time," he says.

Regardless of the reason for the upgrade, Trofimov is a proponent of recyclers reaching out to the technology supplier and working together for the solution. "Recyclers have the knowledge about their operations, what is taken in, what can be sorted, and where improvements are needed," he concludes. "Technology suppliers like TOMRA have vast application experience with machines operating around the world. We have the consultative expertise and will focus on the technical solution to get the right sensor combination to meet established targets."

This article originally appeared in the September/October 2025 issue of Recycling Product News.