Machine learning to predict biochar and bio-oil yields from co-pyrolysis of biomass and plastics

Research output: Contribution to journalArticlepeer-review

43 Scopus citations


Because of high oxygen content, pH and viscosity, pyrolysis bio-oil is of low quality. Upgrading bio-oil can be achieved by co-pyrolysis of biomass with waste plastics, and it is seen as a promising measure for mitigating waste. In this work, machine learning models were developed to predict yields from the co-pyrolysis of biomass and plastics. Classical machine learning and neural network algorithms were trained with datasets, acquired for biochar and bio-oil yields, with cross-validation and hyperparameters. XGBoost predicted biochar yield with an RMSE of 1.77 and R2 of 0.96, and the dense neural network was able to predict the bio-oil yield with an RMSE of 2.6 and R2 of 0.96. The SHapley Additive exPlanations analysis technique was used to understand the influence of various parameters on the yields from co-pyrolysis. This study provides valuable insights to understand the co-pyrolysis of biomass and plastics, and it opens the way for further improvements.
Original languageEnglish (US)
Pages (from-to)125303
StatePublished - Jul 25 2022

Bibliographical note

KAUST Repository Item: Exported on 2022-09-14
Acknowledged KAUST grant number(s): OSR-2019-CRG7-4077
Acknowledgements: This work was supported by King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research under the award number OSR-2019-CRG7-4077. The authors would also like to thank Ibex, high-performance computing at KAUST.

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Organic Chemistry
  • General Chemical Engineering
  • Fuel Technology


Dive into the research topics of 'Machine learning to predict biochar and bio-oil yields from co-pyrolysis of biomass and plastics'. Together they form a unique fingerprint.

Cite this