Optimizing multi-step kinetic schemes for biomass pyrolysis in bioenergy production using multi-objective genetic algorithm: Transitioning from thermally thin to thermally thick regime

This study aims to optimize multi-step kinetic models for biomass pyrolysis to enhance bioenergy production. Utilizing three representative biomass types, i.e., Cunninghamia lanceolata (forestry-derived), rice straw (agricultural-derived), and green algae (aquatic-derived), thermogravimetric analysis was conducted across temperatures from 150 to 600 °C at heating rates of 5, 20, 30, and 40 °C/min. Three kinetic schemes with three, four, and nine reactions were parameterized using a parallelized multi-objective genetic algorithm fitting the model simultaneously to all heating rates. The four-reaction scheme achieved the highest fit quality (R² > 0.97 for all feedstocks), while the nine-reaction model provided detailed mechanistic insights but showed reduced accuracy for green algae (R² as low as 0.81). Coupling these kinetic models with a 2D heat transfer model of a thermally thick biomass particle revealed distinct thermal gradients and reaction front behaviors for each scheme, including average front propagation speeds of 0.19 mm/min and core conversion times between 27 and 34 min. The most complex scheme exhibited the fastest core conversion rate (up to 23 % faster), broadest and most diffuse reaction front, and greater char production. Whereas, simpler schemes produced sharper, localized fronts with slower conversion. This integrated experimental-computational approach quantitatively characterizes pyrolysis kinetics and thermal dynamics, advancing model development for optimized industrial bioenergy applications.