Plastic circular economy framework using hybrid machine learning and pinch analysis

Chin, H.H., Varbanov, P.S., You, F., Sher, F. ORCID: 0000-0003-2890-5912 and Klemeš, J.J., 2022. Plastic circular economy framework using hybrid machine learning and pinch analysis. Resources, Conservation and Recycling, 184: 106387. ISSN 0921-3449

1553947_Sher.pdf - Post-print

Download (1MB) | Preview


The worldwide plastic waste accumulation has posed probably irreversible harm to the environment, and the main dilemma for this global issue is: How to define the waste quality grading system to maximise plastic recyclability? This work reports a machine learning approach to evaluating the recyclability of plastic waste by categorising the quality trends of the contained polymers with auxiliary materials. The result reveals the hierarchical resource quality grades predictors that restrict the mapping of the waste sources to the demands. The Pinch Analysis framework is then applied using the quality clusters to maximise plastic recyclability. The method identifies a Pinch Point – the ideal waste quality level that limits the plastic recycling rate in the system. The novel concept is applied to a problem with different polymer types and properties. The results show the maximum recycling rate for the case study to be 38 % for PET, 100 % for PE and 92 % for PP based on the optimal number of clusters identified. Trends of environmental impacts with different plastic recyclability and footprints of recycled plastic are also compared.

Item Type: Journal article
Publication Title: Resources, Conservation and Recycling
Creators: Chin, H.H., Varbanov, P.S., You, F., Sher, F. and Klemeš, J.J.
Publisher: Elsevier BV
Date: September 2022
Volume: 184
ISSN: 0921-3449
S0921344922002312Publisher Item Identifier
Divisions: Schools > School of Science and Technology
Record created by: Linda Sullivan
Date Added: 27 Jun 2022 15:13
Last Modified: 05 May 2023 03:00

Actions (login required)

Edit View Edit View


Views per month over past year


Downloads per month over past year