Improvement of the Higher Heating Value Prediction of Biomass based on Proximate Analysis using Regression and Neural Network

18 Nov
Posted on Author

Authors: Se Ung Kim, Jong Ryong An, Un Dok Kim, Won Il Ri, Yong Nam Kim

Abstract: – Based on proximate analysis, A study on a regression model to predict the higher heating value (HHV) of biomass was performed. The aim of study is to create the HHV prediction model by regression, thereby improve its prediction performance through modeling on new samples of biomass with wide range in contents of components, and to confirm the result through comparison this model with previous models. The regression model HHV = 7.9457 – 0.0477 × ASH + 0.2839 × FC + 9.7416 E-4 × VM^2 was created by the standard least squares (SLS) method and in result the coefficient of determination (R2), adjusted R2 and the root-mean-square error (RMSE) showed excellent fitness with 0.990, 0.989, and 1.578 for samples with distribution ranges of FC, VM and ASH of 3.86-91.5%, 6.6-87.8% and 0.2-70%, respectively. Sensitivity analysis of the HHV prediction with the number of neurons in the neural network (NN) model showed the best results when the number of neurons was 5.

DOI: https://doi.org/10.5281/zenodo.17639070