Computer-aided energy prediction for selected and blended wood biomass using ultimate and proximate analysis

This study evaluates the energy potential of wood biomass (sawdust) by employing computer-aided techniques to predict the higher heating value (HHV) through ultimate and proximate analyses. The ultimate analysis focuses on elemental properties, while the proximate analysis examines physical properties. The developed regression model demonstrates a high coefficient of determination (R²) of 99.69% for ultimate analysis, indicating a strong predictive capability. In contrast, the proximate analysis reveals individual correlation coefficients of 85.80% for moisture content, 79.18% for fixed carbon, and 28.10% for volatile matter. To assess the significance of each independent variable in the model, the p-values associated with the coefficients were examined. For the ultimate analysis, all input variables except for sulfur (%S) (p ≈ 0.22) had p-values less than 0.05 at a 95% significance level, indicating their statistical significance. However, in the proximate analysis, only volatile matter exhibited a relatively high p-value (p ≈ 0.12), rendering it statistically insignificant in the model. The elevated p-values for sulfur and volatile matter suggest their minimal impact on HHV predictions in their respective models. The computer program developed for this study automates the prediction process, achieving an accuracy within ±5 MJ/kg between predicted and experimental values across the dataset and is uniformly applicable to all individual models or biomass type, significantly reducing analysis time. The findings of this study contribute to optimizing biomass energy systems, enhancing energy recovery efficiency, and advancing sustainable energy practices.