Analysis of biomass milling literature data and energy-particle size laws

This study compiles and analyses a large database of biomass milling energy consumption, comprising 792 experiments extracted from 33 scientific studies covering 59 biomass types and 21 milling technologies. Specific energy consumption (SEC) is examined across coarse, intermediate, and fine comminution regimes, distinguishing net SEC (excluding no-load losses) from total SEC. Results show that SEC increases nonlinearly with decreasing particle size, with fine milling systematically dominating energy demand and exhibiting the largest variability. A feature selection framework identifies biomass structural properties, particularly moisture content, lignin content, and cellulose content, together with target particle size as the primary determinants of net SEC, while machine operating parameters play a secondary role. Classical comminution laws (Rittinger, Kick, Bond), originally developed for brittle mineral materials, show markedly better performance for net SEC than for total SEC, but their predictive capability remains highly feedstock-dependent. Best fits are observed for fibrous herbaceous residues, while woody biomass deviates significantly, reflecting its anisotropic and viscoelastic fracture behavior. Overall, the results demonstrate that biomass milling energy is governed by coupled material–process interactions that are not fully captured by classical energy–size laws. The study highlights the need for biomass-adapted or hybrid physics–data-driven models and provides a structured dataset that can support future development of AI-based predictive frameworks for energy-efficient biomass processing. • 792 biomass milling experiments compiled from 33 studies (59 biomass types, 21 technologies). • Fine comminution shows the highest SEC and variability. • Classical laws fit net SEC better than total SEC. • SEC mainly driven by biomass structure and particle size. • Dataset supports future hybrid physics–data-driven modelling.