Interpretable prediction of biomass-derived biochar characteristics: combining machine learning with shapley additive explanation

Zongqi Chen, Feng Gong, Jiaming Song and Kai Zhang

Energy, 2025, vol. 335, issue C

Abstract: As a significant carbon-neutral product derived from biomass, pyrolysis biochar is extensively utilized in energy, materials, and agricultural fields. However, biochar characteristics derived datasets are high-dimensional and highly non-linear, making them difficult to analyze with traditional methods. In this study, the char energy recovery rate (CER) was innovatively proposed to comprehensively evaluate the coupled effects of biochar yield and higher heating value (HVV). Four data-driven machine learning models were employed to predict biochar HVV, yield, and CER by coupling biomass properties with pyrolysis parameters. Results showed that the Extreme Gradient Boosting (XGB) model demonstrated superior prediction accuracy, achieving regression coefficients of 0.90 for HHV, 0.92 for yield and 0.87 for CER. Partial dependence plots were employed to elucidate the nonlinear relationships between biomass properties and pyrolysis parameters. Shapley Additive Explanations (SHAP) visualization revealed elemental composition most influenced HHV (46.5 %), while process parameters dominated both yield (71.8 %) and CER (67.3 %). The predictions of the XGB models for three typical biomass cases were strongly verified to possess extremely high accuracy and generalization ability, with the error being less than 0.04 %. A visual biochar multi-characteristic prediction platform was developed, providing insights for optimizing the biochar production.

Keywords: Biomass; Biochar; Machine learning; Prediction; Higher heating value (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544225035194
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:335:y:2025:i:c:s0360544225035194

DOI: 10.1016/j.energy.2025.137877

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().