Predicting fuel properties of torrefied refuse-derived fuel using experimental and regression modeling

• Torrefaction improves the fuel quality of RDF. • HTT and CRMM strongly affect the fuel properties of torrefied RDF components. • HTT–CRMM models accurately predict CSF fuel properties for clean RDF fractions. Refuse-derived fuel (RDF), a processed fraction of municipal solid waste, is increasingly used as an alternative fuel in thermochemical applications. Torrefaction, a mild thermal treatment under inert conditions, has been shown to improve RDF fuel properties by increasing carbon content and calorific value while reducing moisture and volatile matter. However, the heterogeneous composition of RDF presents challenges in predicting its behavior during torrefaction. This study presents a modeling framework for predicting key fuel properties of torrefied RDF using experimental data from nine clean RDF components and their binary mixtures. Nonlinear regression models were developed based on torrefaction temperature (HTT) and carbon-relative molar mass (CRMM), a compositional descriptor reflecting molecular structure. The models demonstrated strong predictive power for carbon content (R 2 = 0.888) and heating values (R 2 = 0.810–0.859), while parameters such as ash, nitrogen, and sulfur content were less predictable. CRMM enabled effective representation of compositional variability, particularly regarding carbonization and energy densification. The proposed models offer a practical tool for optimizing torrefaction-based RDF upgrading and provide a foundation for future studies involving more complex RDF mixtures.