Circular economy Using AI to improve recycled plastics

By combining material analysis with pattern recognition software, the Smartreuse project aims to give plastics processors clearer guidance on how to use recyclates efficiently and at consistent quality.

The plastics industry is facing a major challenge: it must increasingly use recycled materials while delivering the same quality as virgin plastics. This is exactly where the “Smartreuse” project comes in. Funded by the European Union and the Free State of Bavaria, it pursues a clear goal: to process plastic waste in such a way that it can be reliably and efficiently used in new products at a high quality level.

Recyclates are plastics recovered either from production residues or from used products. A distinction is made between “post-industrial recyclates” (PIR), which originate from industrial production waste, and “post-consumer recyclates” (PCR), which come from previously used products such as packaging. Both types are to be used more extensively in the project in order to conserve crude oil and reduce CO₂ emissions. “We want to demonstrate that recycled plastics do not have to be a compromise, but can be a genuine alternative to virgin materials,” explains Prof. Dr. habil. Michael Nase, project manager and head of the Institute for Circular Economy of Bio:Polymers at Hof University of Applied Sciences (IBP).

Data analysis instead of lengthy testing

In practice, the use of recycled plastics is often challenging because their properties can vary. Sometimes they are particularly strong, sometimes less durable, and sometimes they behave differently under heat than expected. These so-called thermal properties describe how materials respond to temperature changes. Rheological properties indicate how well a plastic flows in its molten state and how easily it can be processed. Physicochemical properties relate, for example, to the material’s composition and structure. In addition, there are mechanical properties such as strength and elongation of the finished component. All these characteristics are systematically analyzed and documented within the project. “The key is to understand the properties of recyclates holistically,” explains Stefanie Rank, research associate at IBP. “We analyze the materials in a structured way in order to make reliable statements about their processability. This is essential for all subsequent steps.”

The resulting material and process data are evaluated using pattern recognition software. “The software identifies relationships between material condition, machine settings, and product properties, and can derive concrete recommendations from this,” explains Maurice Bäcker, research associate at iisys. These recommendations include, for example, the mixing ratio of different plastics or the use of additives and stabilizers. Additives are substances that improve specific properties, such as durability, while stabilizers help ensure that the material ages more slowly when exposed to heat or light.

“Our goal is to provide companies with clear and easy-to-understand recommendations for action,” says Jörg Scheidt, head of the Analytical Information Systems research group at Iisys. “They should be able to decide more quickly which material mix and which machine settings will lead to the desired result.”

Subprojects in cooperation with regional companies

The project is being carried out at Hof University of Applied Sciences through close collaboration between the Institute for Circular Economy of Bio:Polymers, led by Prof. Dr.-Ing. habil. Michael Nase, and the Analytical Information Systems research group at the Institute for Information Systems, led by Prof. Dr. Jörg Scheidt. It is divided into several subprojects, which are implemented in cooperation with various industry partners. These include the recycling company Komrec-Recond, the plastics processing companies BKW Kunststoff, Wikutec, and Kunststofftechnik Schnitzler, the analytical equipment manufacturer Erich Netzsch Holding, and the software company Smartlytic. Through company-specific recycling projects, the developed pattern recognition software is validated and made usable for small and medium-sized enterprises (SMEs).

The project runs from 2025 to 2028. In the first subproject, various streams of post-consumer recycled HDPE packaging plastics (PCR-HDPE) are being investigated. HDPE stands for “high-density polyethylene,” a particularly durable and widely used plastic often found in bottles and canisters. The aim is to process this material so that it can be used to manufacture holders for beer kegs.

(ID:50806493)