Applied machine learning for prediction of waste plastic pyrolysis towards valuable fuel and chemicals production

Yi Cheng; Ecrin Ekici; Güray Yildiz; Yang Yang; Brad Coward; Jiawei Wang

doi:10.1016/j.jaap.2023.105857

Applied machine learning for prediction of waste plastic pyrolysis towards valuable fuel and chemicals production

Yi Cheng, Ecrin Ekici, Güray Yildiz, Yang Yang, Brad Coward, Jiawei Wang^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract

Pyrolysis is a suitable conversion technology to address the severe ecological and environmental hurdles caused by waste plastics' ineffective pre- and/or post-user management and massive landfilling. By using machine learning (ML) algorithms, the present study developed models for predicting the products of continuous and non-catalytically processes for the pyrolysis of waste plastics. Along with different input datasets, four algorithms, including decision tree (DT), artificial neuron network (ANN), support vector machine (SVM), and Gaussian process (GP), were compared to select input variables for the most accurate models. Among these algorithms, the DT model exhibited generalisable and satisfactory accuracy (R² > 0.99) with training data. The dataset with the elemental composition of waste plastics achieved better accuracy than that with the plastic-type for predicting liquid yields. These observations allow the predictions by the data from ultimate analysis when inaccessible to the plastic-type data in unknown plastic wastes. Besides, the combination of ultimate analysis input and the DT model also achieved excellent accuracy in liquid and gas composition predictions.

Original language	English
Article number	105857
Number of pages	10
Journal	Journal of Analytical and Applied Pyrolysis
Volume	169
Early online date	5 Jan 2023
DOIs	https://doi.org/10.1016/j.jaap.2023.105857
Publication status	E-pub ahead of print - 5 Jan 2023

Bibliographical note

Funding Information:
The work was supported by an Institutional Links grant (No. 527641843), under the Turkey partnership. The grant is funded by the UK Department for Business, Energy, and Industrial Strategy together with the Scientific and Technological Research Council of Turkey (TÜBİTAK; ˙Project no. 119N302) and delivered by the British Council. The author Yi Cheng and Jiawei Wang would like to acknowledge the Marie Skłodowska Curie Actions Fellowships by The European Research Executive Agency (H2020-MSCA-IF-2020, no. 101025906). The author Jiawei Wang would also like to acknowledge the support from Guangdong Science and Technology Program, No. 2021A0505030008.

Copyright © 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

Keywords

Decision tree
Machine learning
Pyrolysis
Ultimate analysis
Waste plastics

Access to Document

10.1016/j.jaap.2023.105857Licence: CC BY 4.0

Chengetal_2023_VoR
Copyright © 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
Final published version, 3.75 MBLicence: CC BY 4.0

Cite this

@article{d330ccf50a3a4886956f5b41804b7a62,

title = "Applied machine learning for prediction of waste plastic pyrolysis towards valuable fuel and chemicals production",

abstract = "Pyrolysis is a suitable conversion technology to address the severe ecological and environmental hurdles caused by waste plastics' ineffective pre- and/or post-user management and massive landfilling. By using machine learning (ML) algorithms, the present study developed models for predicting the products of continuous and non-catalytically processes for the pyrolysis of waste plastics. Along with different input datasets, four algorithms, including decision tree (DT), artificial neuron network (ANN), support vector machine (SVM), and Gaussian process (GP), were compared to select input variables for the most accurate models. Among these algorithms, the DT model exhibited generalisable and satisfactory accuracy (R2 > 0.99) with training data. The dataset with the elemental composition of waste plastics achieved better accuracy than that with the plastic-type for predicting liquid yields. These observations allow the predictions by the data from ultimate analysis when inaccessible to the plastic-type data in unknown plastic wastes. Besides, the combination of ultimate analysis input and the DT model also achieved excellent accuracy in liquid and gas composition predictions.",

keywords = "Decision tree, Machine learning, Pyrolysis, Ultimate analysis, Waste plastics",

author = "Yi Cheng and Ecrin Ekici and G{\"u}ray Yildiz and Yang Yang and Brad Coward and Jiawei Wang",

note = "Funding Information: The work was supported by an Institutional Links grant (No. 527641843), under the Turkey partnership. The grant is funded by the UK Department for Business, Energy, and Industrial Strategy together with the Scientific and Technological Research Council of Turkey (T{\"U}BİTAK; ˙Project no. 119N302) and delivered by the British Council. The author Yi Cheng and Jiawei Wang would like to acknowledge the Marie Sk{\l}odowska Curie Actions Fellowships by The European Research Executive Agency (H2020-MSCA-IF-2020, no. 101025906). The author Jiawei Wang would also like to acknowledge the support from Guangdong Science and Technology Program, No. 2021A0505030008. Copyright {\textcopyright} 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)",

year = "2023",

month = jan,

day = "5",

doi = "10.1016/j.jaap.2023.105857",

language = "English",

volume = "169",

journal = "Journal of Analytical and Applied Pyrolysis",

issn = "0165-2370",

publisher = "Elsevier",

}

TY - JOUR

T1 - Applied machine learning for prediction of waste plastic pyrolysis towards valuable fuel and chemicals production

AU - Cheng, Yi

AU - Ekici, Ecrin

AU - Yildiz, Güray

AU - Yang, Yang

AU - Coward, Brad

AU - Wang, Jiawei

N1 - Funding Information: The work was supported by an Institutional Links grant (No. 527641843), under the Turkey partnership. The grant is funded by the UK Department for Business, Energy, and Industrial Strategy together with the Scientific and Technological Research Council of Turkey (TÜBİTAK; ˙Project no. 119N302) and delivered by the British Council. The author Yi Cheng and Jiawei Wang would like to acknowledge the Marie Skłodowska Curie Actions Fellowships by The European Research Executive Agency (H2020-MSCA-IF-2020, no. 101025906). The author Jiawei Wang would also like to acknowledge the support from Guangdong Science and Technology Program, No. 2021A0505030008. Copyright © 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

PY - 2023/1/5

Y1 - 2023/1/5

N2 - Pyrolysis is a suitable conversion technology to address the severe ecological and environmental hurdles caused by waste plastics' ineffective pre- and/or post-user management and massive landfilling. By using machine learning (ML) algorithms, the present study developed models for predicting the products of continuous and non-catalytically processes for the pyrolysis of waste plastics. Along with different input datasets, four algorithms, including decision tree (DT), artificial neuron network (ANN), support vector machine (SVM), and Gaussian process (GP), were compared to select input variables for the most accurate models. Among these algorithms, the DT model exhibited generalisable and satisfactory accuracy (R2 > 0.99) with training data. The dataset with the elemental composition of waste plastics achieved better accuracy than that with the plastic-type for predicting liquid yields. These observations allow the predictions by the data from ultimate analysis when inaccessible to the plastic-type data in unknown plastic wastes. Besides, the combination of ultimate analysis input and the DT model also achieved excellent accuracy in liquid and gas composition predictions.

AB - Pyrolysis is a suitable conversion technology to address the severe ecological and environmental hurdles caused by waste plastics' ineffective pre- and/or post-user management and massive landfilling. By using machine learning (ML) algorithms, the present study developed models for predicting the products of continuous and non-catalytically processes for the pyrolysis of waste plastics. Along with different input datasets, four algorithms, including decision tree (DT), artificial neuron network (ANN), support vector machine (SVM), and Gaussian process (GP), were compared to select input variables for the most accurate models. Among these algorithms, the DT model exhibited generalisable and satisfactory accuracy (R2 > 0.99) with training data. The dataset with the elemental composition of waste plastics achieved better accuracy than that with the plastic-type for predicting liquid yields. These observations allow the predictions by the data from ultimate analysis when inaccessible to the plastic-type data in unknown plastic wastes. Besides, the combination of ultimate analysis input and the DT model also achieved excellent accuracy in liquid and gas composition predictions.

KW - Decision tree

KW - Machine learning

KW - Pyrolysis

KW - Ultimate analysis

KW - Waste plastics

UR - http://www.scopus.com/inward/record.url?scp=85146173173&partnerID=8YFLogxK

UR - https://www.sciencedirect.com/science/article/pii/S0165237023000013?via%3Dihub

U2 - 10.1016/j.jaap.2023.105857

DO - 10.1016/j.jaap.2023.105857

M3 - Article

AN - SCOPUS:85146173173

SN - 0165-2370

VL - 169

JO - Journal of Analytical and Applied Pyrolysis

JF - Journal of Analytical and Applied Pyrolysis

M1 - 105857

ER -

Applied machine learning for prediction of waste plastic pyrolysis towards valuable fuel and chemicals production

Abstract

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this