Empirical Model for Predicting the Combustion Performance of Low-Quality Biomass Pellets Based on Experimental Data

Low‑quality biomass (LQB) is an abundant feedstock for small‑scale heat generation, but its variable composition often leads to unstable combustion and elevated emissions in domestic boilers. This study assesses the combustion behaviour of 18 pelletized LQB feedstocks using a 25-kW boiler operated under controlled conditions. The fuels exhibited wide performance differences, reflected in broad emission ranges and noticeable variation in thermal efficiency. To support predictive evaluation, statistical models were established that link key fuel characteristics with CO, NO_x, PM and efficiency, enabling quantitative estimation of combustion performance across heterogeneous feedstocks. To clarify ash‑related effects, ten representative ashes were characterized using ICP-OES and XRD. Their chemical and phase compositions were assigned to the main Vassilev classes, which corresponded to distinct ash‑forming mineral groups. Ca‑rich ashes were dominated by carbonate and phosphate phases suitable for neutralization or construction applications. K‑rich ashes contained soluble sulphate and chloride species relevant for nutrient‑recovery routes, while Si‑rich ashes consisted mainly of quartz or amorphous matrices with low reactivity. The combined combustion and ash elemental and phase analysis results demonstrate clear links between LQB composition, emission behaviour and ash transformation mechanisms, providing a practical basis for evaluating the suitability of alternative biomass pellets for small‑scale heat production.