Syngas production through forest waste gasification and prediction of its species using advanced novel metaheuristic driven hybrid machine learning algorithms

Abstract

The present study aims to produce syngas from forest waste through the gasification process. Through this process, significant operation parameters (gasification temperature from 650 to 1000 degrees C with an increment of 50 degrees C, gasifier pressure (1, 3, 5, and 7 bar), equivalence ratio (ER) (0.1, 0.2, 0.3, 0.4, and 0.7)) are varied to observe the changes on hydrogen (H-2), carbon monoxide (CO), and carbon dioxide (CO2) concentrations. Furthermore, this work proposes single-layer hybrid ANN-THRO, ANN-BSLO, and ANN-QIO predictive models to forecast H-2, CO, and CO2 concentrations within the syngas. The results showed that increasing the gasification temperature and reducing the gasifier pressure led to higher concentrations of H-2 and CO in the syngas. However, the ER exhibited a bell-shaped trend for these gases. On the other hand, the CO2 formation generally exhibits an adverse trend against H-2 and CO. In terms of the prediction results, the proposed three hybrid algorithms showed a very satisfactory prediction performance of significant syngas species with an R-2 value of >0.96, rRMSE value of <6 %, and MAPE value of <5.3 %.