Experimental and artificial intelligence study on catalytic reforming of tar over bio-char surface coupled with hydrogen production

Abstract

Tar problem is an obstacle in biomass gasification. Naphthalene was used as tar model compound to investigate the tar catalytic removal over char bed coupled with hydrogen production. The influence of char properties, residence time and atmosphere on tar reduction was taken into consideration. Results show pinewood char acted as quite good performance for catalytic removal of naphthalene. With the increasing of duration time, deactivation would occur, which brought down the tar conversion efficiency. The addition of H2O could inhibit the carbon deposition and promote hydrogen yield via in-situ gasification. At 800 °C, the naphthalene conversion rate slightly declined to 95.77% even at 182 min under 10% steam atmosphere. The H2 yield was around 8.98 mol/(mole of naphthalene) at initial 12 min. The pinewood char pore analysis results confirmed the inhibition of carbon deposition by steam. Artificial neural network (ANN) models coupled with genetic algorithm (GA) and particle swarm optimization (PSO) were built for prediction of naphthalene conversion and hydrogen yield. The BET surface area, potassium content, penetration time, duration time, temperature and atmosphere were used as input variables. Modeling results show that the PSO-ANN model and relative impact analysis could be effectively used for modeling and analysis of tar catalytic conversion over char bed.