Development of a recycled plastics databank to improve the circularity
Principal Organization
Tohoku University
Research overview
The objective of this research and development project is to analyze and improve the quality of recycled materials by integrating physical property and structural measurements, computation, and data-driven science, and to establish a recycled material quality databank necessary for enhancing recyclability. Specifically, this project focuses mainly on recycled polypropylene (PP) and other recycled plastics.
In addition to physical property and structural measurements, physical properties and structures are predicted through molecular structure modeling, and quality improvement is carried out based on the results. Furthermore, aiming to establish models for producing recycled plastics suitable for automotive and home appliance applications from various recycled materials, namely the “X to Car model” and the “X to Home Appliance model,” quality analysis and quality improvement are conducted for high-quality recycled plastics applicable to automobiles and home appliances.
By the end of the project, a recycled material databank will be established by accumulating 10,000 sets of physical property and structural data. These data will be linked with information on recycled material fabrication requirements related to the plastics information distribution platform, PLA-NETJ, such as the collection area, sorting method and date of waste plastics, as well as the pelletizing date and processing environment. For the operation of the recycled material databank, a service window will be established for receiving recycled material pellets as “deposits” and providing secure measurement data as “withdrawals.”
When providing measurement data to recyclers, data-driven science, including missing-data completion by Bayesian estimation and clustering by self-organizing maps (SOM), will be introduced.
This approach enables the analysis of correlations between physical properties and structures for recycled materials, whose quality varies daily depending on collection area, date, sorting method, and other factors, as well as for improved recycled materials developed through quality improvement studies. Through this analysis, the mechanisms of property degradation in recycled materials will be clarified, and the analytical results will be provided together with their interpretation, namely scientific insights from researchers, which correspond to “interest” in the databank concept.
In addition, measurement data for virgin plastics will also be accumulated in the SOM clustering dataset, making it possible to clarify the relative position of the physical properties and structures of recycled materials compared with those of virgin materials. These analytical data will also be utilized for grading recycled materials.