Predictive Modelling of Hydrogen Production from Agricultural and Forestry Residues through a Thermo-catalytic Reforming Process

Hydrogen produced from renewable sources is crucial for decarbonizing hard-to-abate sectors and achieving netzero targets. This study examines hydrogen production through the novel thermo-catalytic reforming (TCR) process using agricultural and forestry residues. The research aims to develop and optimize regression models that integrate feedstock properties (ash, hydrogen-to-carbon molar ratio, and lignin) and process parameters (reactor and reformer temperatures) to predict yields of hydrogen (H2), syngas, methane (CH4) and carbon dioxide (CO2). Three biomass feedstocks – softwood pellets (SWPs), hardwood pellets (HWPs), and wheat straw pellets (WSPs) – were analyzed at reactor temperatures of 400–550 ◦C and reformer temperatures of 500–700 ◦C. Predictive models for H2 (R2 = 0.9642, RMSE = 1.0639) and syngas (R2 = 0.9894, RMSE = 0.0140) yields show strong agreement and accuracy between the predicted and experimental values. In contrast, the models for CH4 and CO2 yields show higher variability in the predictions. Reformer temperature was the most significant parameter influencing the yields of H2 and syngas. The optimal H2 yields predicted for the model were obtained for HWPs at 550/700 ◦C (26.67 g H2/kg dry biomass), followed by SWPs at 550/700 ◦C (24.11 g H2/kg dry biomass) and WSPs at 550/685.2 ◦C (18.78 g H2/kg dry biomass). The volumetric syngas yields were highest for HWPs at 550/700 ◦C (0.831 Nm3 /kg dry biomass), followed by SWPs (0.777 Nm3 /kg dry biomass) and WSPs (0.634 Nm3 /kg dry biomass). This study demonstrates that regression modelling accurately predicts H2 and syngas yields, which would help to expand the applicability of TCR technology for large-scale hydrogen production, contributing to the decarbonization of the energy sector.