(731d) Modeling Biomass Pyrolysis at the Particle Scale: How Anisotropic Permeability Impacts Conversion Time and Reactor Sizing in Anoxic and Autothermal Conditions

The fast processing time of biomass pyrolysis allows for high throughput conversion of seasonal feedstocks like forest management wood thinnings and agricultural residues. However, scale-up remains more art than engineering due to particle material attributes, non-ideal reactors, and carrier oxygen content impacting the yield and quality of the final oil and char.

Biogenic materials have anisotropic directional porosity for water/nutrient transport from roots to extremities. Many pyrolysis models incorporate anisotropic thermal conductivity and diffusion with non-spherical particles (Pecha, et al. 2018) or equivalent spheres (Gao, et al. 2021). However most do not account for the pressure released inside due to solid-gas phase-change. Yet gas K­_axial/K_radial (permeability) has been measured as high as 1,000,000x higher along the wood grain vs across. A 3-d particle model was constructed with FEM based on compositional and imaging analysis using the CRECK pyrolysis scheme (Debiagi, et al. 2018) for a spruce particle (Anca-Couce, et al. 2017), heated at 550 °C under inert gas (Fig. 1). Neglecting anisotropic porosity attenuated convective HX through the sides of the wood cylinder, more than doubling the measured conversion time (Fig. 2). Permeability tensors of K­_axial/K_radial >10,000 had realistic conversion times. Results were incorporated into a techno-economic analysis (TEA) for pyrolysis to predict bio-oil costs.

In a second effort, anisotropic material attributes were simulated for autothermal pyrolytic conversion wherein oxygen is fed into a pyrolysis reactor to provide heat via oxidation (Polin, et al. 2019). This model included a char oxidation scheme (Evans and Emmons 1977) for oak dowels. Results showed that if transverse mass transport is set 1:1 with axial, advection out of the particle reactions attenuates HX and oxygen diffusion. When K­_axial/K_radial is >10,000, conversion times match experiment (Fig. 3). This model was then used in a reactor scale simulation for autothermal pyrolysis of corn stover.