Unified interpretable machine learning framework for predicting pellet quality from raw and thermochemically pretreated biomass

Muzammil Khan; Xiangpeng Gao; Kok Wai Wong; Shinji Kudo; Jun-Ichiro Hayashi; Hongwei Wu

doi:10.1016/j.biortech.2026.134909

Back

Unified interpretable machine learning framework for predicting pellet quality from raw and thermochemically pretreated biomass

Journal article

Open access

Peer reviewed

Unified interpretable machine learning framework for predicting pellet quality from raw and thermochemically pretreated biomass

Muzammil Khan, Xiangpeng Gao, Kok Wai Wong, Shinji Kudo, Jun-Ichiro Hayashi and Hongwei Wu

Bioresource technology, 134909

2026

DOI: https://doi.org/10.1016/j.biortech.2026.134909

PMID: 42150639

Files and links (1)

pdf

Published15.85 MBDownload View

Published (Version of Record) Open Access CC BY V4.0

Abstract

Thermochemical conversion

Densification

Optimisation

Artificial intelligence

Pelletisation

Highlights • Unified ML model accurately predicts pellet quality for raw and pretreated biomass. • Thermo-mechanical process rules act as key determinants of pellet quality. • Mechanistic insights reveal chemistry-driven and pretreatment-specific limits. • Unified model reduces computation by 25% compared to pathway-specific models. • Graphical user interface enables feedstock-agnostic prediction of pellet quality. Abstract Biomass pellets offer a renewable pathway for decarbonising industrial energy and metallurgical processes, yet inconsistent quality limits widespread adoption. This study presents a unified, interpretable machine learning framework that predicts pellet density and mechanical strength across raw biomass, hydrochar, and torrefied biomass feedstocks. A literature-derived dataset spanning feedstock properties, pretreatment conditions, and densification parameters was compiled to capture heterogeneity in pelletisation systems. Eight algorithms were benchmarked using Bayesian optimisation and 5-fold cross-validation, achieving R2 > 0.85 with root-mean-square errors within experimental uncertainty. Interpretability analyses revealed critical nonlinear interactions among binder content, feedstock composition, and thermo-mechanical conditions that influence densification performance. The unified framework matched experimental-grade pellet specifications (∼1.2 g cm−3; 6–7 MPa) while reducing computational cost by approximately 25% compared to pathway-specific models. The framework provides generalisable thermo-mechanical design principles and is deployed through a graphical interface, enabling users to predict pellet quality based on compositional and process inputs.

Details

Title: Unified interpretable machine learning framework for predicting pellet quality from raw and thermochemically pretreated biomass
Authors/Creators: Muzammil Khan - Murdoch University
Xiangpeng Gao - Murdoch University
Kok Wai Wong - Murdoch University
Shinji Kudo - Kyushu University
Jun-Ichiro Hayashi - Kyushu University
Hongwei Wu - Curtin University
Publication Details: Bioresource technology, 134909
Identifiers: 991005883831507891
Murdoch Affiliation: School of Engineering and Energy; School of Information Technology; Centre for Water, Energy and Waste
Language: English
Resource Type: Journal article

UN Sustainable Development Goals (SDGs)

This output has contributed to the advancement of the following goals:

Source: SDGs in the Output

Metrics

1 File views/ downloads

1 Record Views