Co-pyrolysis of biomass and plastic wastes and application of machine learning for modelling of the process: A comprehensive review

Abstract

The conventional fossil fuels which primarily include coal, oil and natural gas are the major source of greenhouse gas emissions (such as methane, carbon dioxide and nitrous oxide) into the atmosphere causing severe health consequences to human population. Different types of renewable energy feedstocks including biomass wastes are being investigated across the world. Out of various techniques for utilizing biomass, the pyrolysis has wide product profiles which can be used in different applications. Likewise, omnipresence of plastic waste, and its tremendous generation and lack of appropriate waste management system is also another environmental issue. Hence, co-pyrolysis (a thermochemical conversion) of biomass and plastic waste, presents an effective solution for the underlined issues as it not only provides a clean source of energy, but is also cost-efficient, easy to use, helps deal with the issue of plastic waste management as well as mitigate the concerns caused by the pyrolysis of single feedstock i.e., biomass. The quality of co-pyrolysis derived bio-oil can further be enhanced by incorporating catalyst. Operating condition of a pyrolysis process depends on the nature of feedstock, requirement of product distribution etc. Thus, optimization of process parameters is essential for making this process successful. Machine learning models can be utilized in the co-pyrolysis process as a tool to overcome the preceding issues by optimizing the process and also helps in process control, yield prediction and real-time monitoring. However, no prior study has conducted an in-depth review of current research scenario related to the machine learning approach in co-pyrolysis process.