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

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 R-2 of 0.96, and the dense neural network was able to predict the bio-oil yield with an RMSE of 2.6 and R-2 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.