Enhancing plastic pyrolysis for carbon nanotubes synthesis through machine learning integration: A review

Abstract

The extensive use of plastic in daily life poses significant risks to the environment in recent years. Thus, transforming plastic waste into a valuable resource is essential to accomplish Sustainable Development Goal (SDG) 11: Sustainable Cities and Communities and SDG 12: Responsible Consumption and Production. The pyrolysis process as an alternative solution is proposed for the conversion of plastic waste into products such as oils, gases, and chars. A search on the Scopus database yielded 355 studies since 2021 on "pyrolysis using machine learning (ML)". This study highlights the application of ML techniques to enhance predictive modelling in plastic pyrolysis processes. ML has been effectively implemented in various aspects of plastic pyrolysis. However, its application in the production of carbon nanotubes (CNTs) has still not been fully explored. Various ML methods such as regression, decision trees, artificial neural networks (ANN), support vector machines (SVM), Random Forest (RF), Gradient Boosting Regressor (GBR), eXtreme Gradient Boosting (XGBoost), AdaBoost, and Stochastic Gradient Descent (SGD) are reviewed in terms of their effectiveness in predicting pyrolysis outcomes, considering different feedstock types, dataset sizes, and input/output variables. The study also addresses current challenges and prospects, focusing on the production of CNTs from plastic waste and optimizing pyrolysis conditions. In conclusion, integrating ML into plastic pyrolysis processes enhances both process efficiency and economic viability.