Development of Calorific Value Prediction Model Using Torrefied Biomass Elemental, Proximate and Chemical Composition

초록

With increasing environmental degradation, biofuels such as biochar have emerged as viable alternatives to fossil fuels. The heating value of biofuels is a key parameter for assessing their efficiency, yet it is often lower than that of traditional fossil fuels. Direct measurement of heating value requires multiple analyses, consuming both time and resources. To address this, numerous predictive models based on various analytical approaches have been developed. This study constructed heating value prediction models using elemental analysis, proximate analysis, and chemical composition analysis, applying them to thermally treated herbaceous agricultural biomass (soybean pods, corncobs), woody agricultural biomass (pepper stems, grape pruning branches), and forestry biomass. A dataset of 66 samples was analyzed. Model performance was evaluated using the coefficient of determination (R2) and root mean square error (RMSE), with elemental analysis models achieving the highest accuracy (R2 = 0.9302 - 0.9650), followed by chemical composition (R2 = 0.8920 - 0.9082) and proximate analysis (R2 = 0.7882 - 0.9882). The optimal models were selected based on both statistical performance and agreement with previous literature. These equations can support rapid estimation of heating values, reducing the need for direct calorimetry. The findings have potential applications in industrial bioenergy production, agricultural residue management, and biomass-based energy policy planning.

키워드