The purpose of this study is to analyze the reliability of predictive models for higher heating values related to organic materials. A theoretical model was developed, which utilizes bond dissociation energies (BDEs) to establish correlations between elemental composition and calorific values. Our analysis indicates that the energy contribution of one mole of hydrogen atoms is approximately equal to -144.4 kJ mol-1. Further investigation reveals significant variations in the bond dissociation energies of carbon atoms within organic compounds, resulting in a range of energy outputs from -414.30 to -275.34 kJ mol-1 per mole of carbon atoms. The presence of oxygen atoms in organic compounds has a negative impact on the magnitude of combustion heat, with values ranging from 131.1 to 207.17 kJ mol-1. The combustion mechanism imposes certain constraints, leading to the equation HHVg = -31.34·[C] - 144.44·[H] + 10.57·[O] for organic compounds. Based on the parameter sensitivity analysis, the coefficient associated with carbon mass fraction exhibits a significantly greater impact on result prediction accuracy, demonstrating a sensitivity value of 92.65%. The results of further analysis indicate that empirical correlations involving the mass fractions of the elements N and S in lignocellulosic materials may be prone to over-fitting, with sensitivity indices of 1.59% and 0.016%, respectively.
Keywords: bond dissociation energy; heating value; oxygen consumption; reliability analysis.