Machine learning predicting and engineering the yield, N content, and specific surface area of biochar derived from pyrolysis of biomass

Abstract

Highlights Yield, N content, and specific surface area of biochar predicted by machine learning Gradient boosting regression outperformed random forest, with test R2 of 0.81–0.95 Temperature, nitrogen, and ash were top features for predicting the three targets The yield and properties of biochar were engineered and experimentally verified

Keywords

• Specific surface area
• Nitrogen
• Biochar
• Pyrolysis
• Machine learning