Dirty MRFs and Clean Products

What separates winners from losers in a tight market? Since China declined to take further bulk quantities of recyclable materials from the US last year, unless they met new very stringent quality standards, the recycling market has experienced a downturn with the loss of its single largest market. Something similar happened to the recycling industry after the stock market crash and recession of 2008. Though in this earlier case it was a general dropoff in demand for recycled materials due to a fall in economic activity, in both cases the demand and price for recyclables were negatively impacted. And in both cases, the market for recyclable materials has and will recover.

While the recycling industry is working to find other markets, the industry winners will be those who increase productivity, improve quality, and cut prices. And this will depend on advances in sorting technology and concurrent improvements and innovations. At the forefront will be advances in sensor technology and optical sensing to achieve the necessary purity levels demanded by improving quality. And although advanced AI and robotics may one day create a “Smart MRF,” the human element will always remain paramount, requiring ever improved levels of training and intelligent, market-focused planning.

Multi-Stream and Single-Stream MRFs
Single-stream MRFs are the opposite of multi-stream MRFs in every important operational characteristic. Instead of receiving multiple streams of waste from various sources and locations, single-stream MRFs receive waste directly from a single source, the community’s waste collection operations. As such, the commingled material that arrives at a single-stream MRF is far from pure, which give the facility its other name of “dirty MRF”. This type of facility relies primarily on machines to perform its sorting and separation operations. Waste arrives at the MRF’s tipping floor and is loaded onto a conveyor belt which carries the wastestream through various removal stations. Each of these stations is designed to remove a particular type of material from the waste. These machines perform their operations based on the size, electromagnetic properties, shape, weight, color, and density of the material being removed. The primary types of machines and the materials they remove are as follows:

• Magnetic separators for removal of ferrous metals: This is a simple and straightforward mechanism that relies upon electromagnetism to remove steel and other ferrous metals directly from the wastestream. What varies is the configuration of the magnets and the adjacent conveyor belt carrying the wastestream. The magnets either can be set overhead, pulling up the ferrous objects out of the wastestream passing underneath on the conveyor belt, or can be integrated into the belt itself, causing the ferrous metals to stick to the belt while the rest of the wastestream falls off into bins or another conveyor belt carrying the waste to the next removal station. The belts and the attached ferrous metals turn under at a roller located at the end of the conveyor belt. There, the metal is scraped from the belt by an edged blade and drops off into a designated collection bin.
• Eddy-current separators for the removal of non-ferrous metals: This is a more complex system than an electromagnet, but it is required to extract the bulk of the metals in the wastestream that are not ferrous. Its operation is based on the method of induced currents and the generation of an electromagnetic field from these currents. A collection of fixed magnets is arranged around the rim of rapidly spinning rotors. As the rotor spins, its magnets induce an electrical current in each piece of non-ferrous metal. This current, in turn, generates its own electromagnetic field in opposition to the field created by the fixed magnets. The two fields repel each other and the non-ferrous metal literally leaps off of the belt into a waiting receptacle.
• Disc screens for the removal of OCC: Disc screens are used to remove large but light objects (such as OCC boxes or sheets). It does so by generating a wave action in the incoming waste. This wave action carries the larger and lighter objects to the top of the wastestream for easy removal. The wave is generated by a series of rotating discs of various shapes and sizes (circular, oval, star-shaped, etc.) set in parallel and intermeshing rows of the disc screen floor. The wastestream enters the floor area as agitated by the rotating discs. Heavier and small objects pass through the screener and on to the next removal stage.
• Rotating trommels for the removal of small contaminant particles: Trommels remove the wastestream’s smaller and heavier particles. Originally designed for mining operations to separate slag from ore, trommels are large metal drums spinning on a rotating axis that is set at an angle to the vertical. Inside the drum are a series of parallels spiraling vanes that cause the wastestream to tumble as the drum spins. The drum itself is perforated with various sized holes. These holes allow the smaller and heavier particles (fines, grit, shards, and organics) further removing impurities from the wastestream.
• Air Classifiers for the removal of bulk quantities of paper: This equipment resembles a tall chimney with a large blower installed at the top of the stack to suck the air out of the classifier at high velocity. The wastestream is fed into the air classifier at its midpoint. Lightweight paper is sucked to the top of the stack while heavier, non-paper objects fall to the bottom and go on to the next recycling stage.
• Air knives for the removal and separation of different types of paper: This is a more refined use of the air classifier that utilizes precise parallel sheets of high-velocity air streams (the air knives) arranged in multiple layers and utilizing different speeds to remove paper objects of differing densities and grades with a high degree of accuracy.
• Glass-color separators for the removal of glass and cullet based on their color: This equipment uses the technology of light spectro-photometry (LSP) to differentiate between the various colors of glass and ceramics. The wavelengths of the different colors reflected back from the glass objects trigger a sensor which determines what color it is. Once the glass color is determined, the sensor triggers an air blower that pushes the glass into an appropriate storage bin.
• Optical sensors utilizing near-infrared (NIR) technology are used to differentiate among both colors and densities to sort different types of plastic. These sensors have the ability to detect and distinguish between the unique reflective signatures of each type of plastic produced by its color and density. Like the glass color separators, the NIR optical sensors trigger air jets that push plastic off of the sorting belt into their respective collection bins.

Some MRFs have additional equipment to handle other types of waste:

• Hydro-pulpers are another means of handling organic materials, instead of source separating them out prior to shipment to the MRF. Used at facilities referred to as “wet MRFs,” hydro-pulpers can also handle lightweight paper products and can be used instead of air classifiers or air knives. Unlike the air knife, there is nothing sophisticated or refined in this brute force method. Jets of high force and high-velocity water are used to pulp paper and organic waste so that it can be drained and removed from the conveyor system. The pulped organic waste is then sent to an anaerobic digester to be turned into biogas. Though effective in producing a high-grade gas fuel, this system requires large amounts of water to operate requiring additional costs and an interior water reclamation system to minimize water supply needs.
• While human laborers should never be thought of as machines, humans remain a primary component of any MRF operation. Humans remain the only component of the MRF operation that can think and learn from experience, finding innovative ways to solve production problems as well as control the operations.

The equipment described above and the material they are designed to separate and sort are summarized in Table 2.

Coming Advances in MRF Technology
The next stage after optic sorting is the expanded use of robotics. And the next stage after that is the use of artificial intelligence that can learn and make decisions in regards to what materials are being sorted and separated from the incoming commingled wastestream. The end goal is a Smart MRF that can theoretically receive a mixed wastestream and proceed to utilize advanced sensors and optics, neural networks, algorithms capable of “learning,” and robotic machinery that never gets tired or bored to efficiently extract individual types of waste and separate them into stockpiles for recycling.

Up until now, even the most that an advanced near-infrared sensor could do is actuate a blast of air that would knock a piece of waste into an adjacent collection bin. The next generation of near-infrared detectors will be integrated with sensors that analyze surfaces in three dimensions along with high-speed cameras with high resolution that will not only be able to distinguish waste materials, but will also be able to recognize a milk bottle from a soup can.

Instead of a low-efficiency, potentially hit-or-miss puff of air, these next-generation sensors will operate robotic armatures that will physically pick through the wastestream at high speeds, removing each different type of waste one-by-one until nothing is left at the end of the conveyor belt. And as they remove each item, the object will be evaluated for size, shape, density, and even market value—all in real time.

And it’s the robots that have some catching up to do. Optical sorters can evaluate waste objects at a rate 10 times faster than current robots can physically remove them. It takes about a tenth of a second for an optical sensor to analyze and categorize a waste object. Meanwhile, even the most advanced robot system takes a full second for you to analyze an object, grab the object, and dispose of it. It is not just the physical motion that takes up time; the robot is using sophisticated algorithms that allow it to learn and improve its performance with time and experience. In effect, the robot benefits from “OJT.”

The system utilized by the more advanced version of waste-picking robots is similar to that of facial recognition programs. The same technology that can be used to pick out a terrorist from a crowd of people can be used to differentiate between a paper cup and a plastic cup of similar shape and size. These machines have neural networks that use the process of “deep learning” to build up a storehouse of knowledge in regards to waste materials. Once these shapes, dimensions, and images are stored, the optical sensors and cameras can send their images of the object on the conveyor belt to the robot’s neural network. Millions of such images, shapes, and material specifications can be stored in the robot’s neural network. The robot can then decide if the object is a plastic shampoo bottle and can be removed from the wastestream or a broken coffee cup which should be left to the ceramic removing robot further down the line. Advanced waste sorting robots will also have the ability to prioritize and pick the most valuable object at hand. The robots can even speak to each other, with robots at the start of the conveyor sending information down the line to other robots, letting them know that the material they should extract will be coming soon.

And the robots will be able to learn. After performing millions of waste removal and evaluation tasks, the robot will be able to judge completely new objects, shapes, and configurations. With this learning, the robot can then correctly judge the characteristics of these never-before-seen objects and decide to sort or not. And this learning will allow the robot to make a decision based on probability instead of just a simple yes/no decision. It can evaluate the object based on how similar it is to other previously sorted objects and correctly identify it with a high degree of accuracy.

In addition to being able to learn, MRF robots will be able to give themselves checkups. Maintenance-related self-diagnostics is a standard feature of industrial robots and it’s only a matter of time before these systems are applied to MRFs. This will allow robots to self-evaluate wear and tear from standard operations, estimate the remaining operational lifetimes for key parts and components, schedule regular maintenance, and report the need for any repairs. Auto-calibration will allow the MRF robot to adjust its operations to maintain high performance without the need to stop operations and perform manual calibrations.

Additionally, the entire MRF will be able to self-evaluate its own operational performance, productivity, and profitability, and can make changes as needed to meet the requirements of new, changing, or expanded wastestreams—and changing market demands. This includes assessments of the final products of the MRF operation, especially the product’s purity and its lack of commingled contaminant material. This is of vital importance. Major consumers and markets for recyclables (like China) have established far stricter standards for recyclable material quality.

Emerging hand in hand with smart robots capable of learning will be ever more sophisticated optical systems that allow them to see. These systems will integrate color analyses, spectral evaluations to determine material makeup, and object recognition too. So, an optical sensor can distinguish between a piece of black plastic HDPE pipe and a clear plastic milk jug.

Major Suppliers
Cambridge Companies is a family-owned design-build ­construction firm with offices in Northwest Indiana and Scottsdale, AZ. They serve clients who desire a customized design-build solution. Their company’s core services include design engineering, construction management, and other services. Cambridge was founded in 1988 by Ray Eriks to develop premium heavy commercial and light industrial design-build projects, originally in the NW Indiana and NE Illinois areas, but they have expanded their business to serve the waste industry nationally. Jeff Eriks is Vice President of Business Development and Marketing and Evan Williams is Design Project Manager at Cambridge Companies. Cambridge is founded on the principles of exceptional client services and delivering a top-quality project on time and within budget. As a design-build firm working with the waste industry for more than 30 years, Cambridge remains an industry leader. During this time, they have completed more than 100 solid waste design-build projects including new builds, repairs, upgrades, and/or modifications at transfer stations, recycling centers/MRFs, hauling companies, landfill facilities, office buildings, and more. Major projects completed by ­Cambridge include:

• Southern Nevada Recycling ­Center. In order to meet the increasing demand for processing recyclables in Nevada, the client, equipment manufacturer, and Cambridge worked together to develop an efficient, durable, long-term solution for the market. The new recycling center incorporates state-of-the-art electrical components and control systems, natural daylighting, safe workflow for the employees, and solar panels on the roof. In order to achieve flexibility, the building was designed to be expandable in multiple directions to add more square footage under roof. The new facility will serve the client and the greater Las Vegas market for years to come.
• Jacksonville Recycling Center. Cambridge was tasked by our customer to develop a recycling facility to serve their marketplace that was going to experience a sharp increase in recycling volumes. After evaluating several existing building options, it was decided that the best solution would be to design and build a new building on an existing piece of property they owned. This resulted in a 73,000-square-foot, state-of-the-art, durable recycling facility that could process up to 35 TPH today while also being designed to accept a second line to double capacity without shutting down the existing line. At any point in time, the owner can add the second line with minimal disruption to their current operation in order to double their throughput. The third floor education center allows the owner to bring in groups of up to 60 people with full visibility to the process area as well as camera views throughout the facility. The end result was a facility that exceeded the customer’s expectation and will serve them for years to come.
• SOCCRA MRF. Southeastern Ocean County Resource Recovery Authority (SOCRRA) operates a Transfer Station and MRF in Troy, MI, that was not able to keep up with the increasing demands of the participating communities. Machinex and Cambridge were brought on to expand the MRF facility. This was a challenging project due to the land-locked limits of property space. Together, Cambridge and Machinex designed and built an expansion of the existing building, new tipping floor, and state-of-the-art automated recycling sorting machinery that can now handle 100 tons of materials per eight-hour shift.