Multiphysics modeling and simulation of high-solids dilute-acid pretreatment of corn stover in a steam-explosion reactor

Abstract

Pretreatment of lignocellulosic biomass, by which the accessibility of cellulose to enzymes is enhanced, is an important step in the biochemical conversion of biomass feedstocks to biofuels. A mathematical model that considers both spatial and temporal phenomena in a biomass particle has the potential to accurately predict carbohydrate conversion and macroscopic structural changes during pretreatment. The development, implementation, and validation of a multiphysics model for high-solids (44%) dilute-acid pretreatment of a pre-impregnated corn-stover particle is presented. The model consists of tightly coupled time-dependent reaction-diffusion equations with finite-rate chemistry. The model partial-differential equations were discretized and solved numerically by the Legendre spectral finite element method. The simulation results were compared to experimental data that were obtained by performing pretreatment on pre-impregnated corn stover in a steam-explosion reactor. The majority of the modeling parameters were set by established science, obtained from the literature, or were experimentally measured initial conditions. The remaining unknown kinetic rate parameters were obtained by fitting simulation results to a limited set of experimental data. Reasonable quantitative agreement was subsequently obtained between simulation results and a larger set of experimental data. The model simulations illustrate the strong dependence of the biomass conversion rate on the spatially dependent and transient acid concentration within the particle. It is observed that time scales of thermal-diffusion are much shorter than those for chemical reactions, which results in an isothermal condition for long pretreatment times. Diffusion of steam within the biomass particle dictates acid dilution and the zone where reactions occur.