Yapay Zeka Hızlı Bakış & Çıkarımlar
AI InsightHowever, the improvements created by non-linear equations were negligible and it was concluded that the linear empirical equations provide satisfactory prediction performance and they may be tried to estimate the calorific value of very wide range of biomasses.
📌 3 Maddede Bu Makale
Abstract
Thirty nine different biomass samples ranging from various herbaceous/woody materials to juice pulps were used to develop linear as well as non-linear empirical equations that predict the lower heating value (LHV) and the higher heating value (HHV) based on the elemental analysis (C, H, N, O, and S) results of the biomass species. These equations were interpreted with respect to their prediction performance considering the predicted values and the experimental data. Several criteria such as mean absolute error (MAE), average absolute error (AAE), average bias error (ABE), and root mean square deviation (RMSD) were regarded. For the linear equations, it was found that the lowest values of MAE were 0.3119 MJ/kg and 0.2906 MJ/kg for HHV and LHV, respectively, and AAE(%) changed in the ranges of (1.6659-4.5917) for HHV and (1.8216-5.5039) for LHV. Besides, it was determined that ABE(%) varies in the intervals of (0.0549-0.2976) for HHV and (0.0519-0.4177) for LHV when linear equations were tested. The best results of RMSD (0.4230 and 0.3607 for HHV and LHV, respectively) were obtained for Equation#1 where all of the linear terms were considered. Also, the addition of the non-linear terms to the linear equations was also studied to check whether any further improvement can be achieved in predictions. However, the improvements created by non-linear equations were negligible and it was concluded that the linear empirical equations provide satisfactory prediction performance and they may be tried to estimate the calorific value of very wide range of biomasses.