10105294Dynamic Optimisation of Manufacturing Parameters for the Processing of Low-Grade, Mixed Plastic Waste as a FeedstockClosedCollaborative R&DInnovate UKThis project will use machine learning to dynamically optimise the manufacturing parameters of a plastic extrusion line in response to variable post-consumer input streams. This will enable manufacturers to standardise the output from mixed, low-grade polymer feedstocks, which make up 30-50% of the throughput of recycling facilities, and virtually all of which currently go to waste. By standardising the output, this technology will remove the main obstacle to the uptake of recycled materials within the polymer manufacturing sector. This will dramatically increase the efficiency of plastic recycling, with the potential to increase national plastic recycling rates by 25-30%. Currently, only a small percentage of recycled plastic can be used by manufacturers because of quality control issues. Manufacturers need precise and consistent levels of mechanical performance, UV stability, and processing viscosity in their feedstock, but most recycled plastic exhibits large variation in these parameters because they cannot be fully purified into individual polymer types and therefore incorporate residual mixtures. Even those that can be purified are often made from comingled sources that are made up of different grades and therefore have variable performance, meaning that only low levels of recycled content can be used. This project will make build on a prior InnovateUK-funded project that enhanced low-grade plastic streams. It successfully proved that mixed polymer streams could, in theory, achieve consistent mechanical performance. However, different polymer streams require different processing conditions, and we discovered that changes in input require a change in temperature profile, screw rotation rate, processing additives, and cooling time. These conditions are normally arrived at through trial and error, but a more precise and efficient way of optimising the line for recycled material is needed for variable post-consumer streams. This project will enable these parameters to dynamically respond to changes in the input stream, enabling the efficient use of variable mixed streams as a feedstock for the construction materials market. Doing so will enable Circular11 to produce certification-grade products out of plastic waste, validating the approach before scaling it into the manufacturing sector itself.