An Artificial Intelligence-Based Modification to NHV Estimation of Biomass

초록

This paper presented an artificial intelligence-based modification to the net heating value (NHV) estimation of biomass from their proximate analyses data earlier proposed in the literature. Artificial neural network (ANN) and gene expression programming (GEP) were used for the modification. The moisture content (M), fixed carbon content (FC), ash content (A), volatile matter (VM), and organic matter (OM) are the input parameters and the NHV is the expected outcome. The ANN-based model (M1) was developed using the weights and biases obtained from the ANN simulation, while the GEP model (M2) directly presents its result in the form of a mathematical model. The feat of the proposed models was also tested using the adjusted coefficient of determination (Adj. R2), mean absolute percentage error (MAPE), mean absolute error (MAE), and variance accounted for (VAF) as the performance indices. The performance of the proposed models is better than the models subjected to modification as the proposed models have adj. R2 and VAF values of 0.962 and 96.17% for the M1 and 0.912 and 91.68% for the M2, while the original model has adj. R2 and VAF values of 0.88 and 88.41%. Hence, the proposed models have significantly improved the model subjected to modification. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

키워드