Fast pyrolysis simulation via kinetic approach and multivariate analysis to assess the effect of biomass properties on product yields, properties, and pyrolyzer performance

Fast pyrolysis (FP) is the thermochemical conversion of biomasses and other carbonaceous materials into bio-oil (liquid), char (solid), and gas products to be used to generate heat, power, chemicals, and fuels of lower greenhouse gas emissions. Considering biomass properties highly affect the FP performance, this work used simulation and multivariate analysis tools to assess the effect of the composition of nine Brazilian biomasses on the FP energy demands, product yields, and bio-oil properties. A simulation, representing a pyrolyzer (480 °C, 1-s residence time), product recovery, and char combustion to meet FP energy demands, was built in Aspen Plus v.10 and validated against experimental data. A dataset correlating biomass compositions from 60 sources and FP outputs was obtained and analyzed, via two statistical methods, to detect similarities between feedstocks and correlate biomass inputs and process outputs. Hierarchical cluster analysis (HCA) separated the feedstocks into two main clusters: agricultural and woody biomasses. The main differences were associated with agricultural feedstocks having higher biomass H/C ratios, lower carbon and volatiles contents, and being of the CHL type (cellulose > hemicellulose > lignin), resulting in lower char and higher gas yields. As for correlations between inputs and outputs, HCA and principal component analysis (PCA) divided the parameters into five main groups: (i) biomass O/C ratio and cellulose content; (ii) bio-oil yield, heating value, and extractives content; (iii) biomass H/C ratio, hemicellulose content, and bio-oil quality metrics; (iv) ash content, char yield, and pyrolyzer heat duty; and (v) biomass lignin, volatiles, and heating value with gas yield and heating value. Several anticorrelations were also detected (e.g., bio-oil yield anticorrelated to bio-oil properties; char yield anticorrelated to gas yield). This work hopes to serve as a simplified roadmap for relationships between biomass properties and FP performance, providing guidelines for biomass selection for different bio-oil downstream applications and perspectives on future works on FP simulation.