Chemistry-Informed Neural Networks modelling of lignocellulosic biomass pyrolysis.

Biomass pyrolysis is a complicated reaction process that involves complex components and reaction pathways. Due to measurement limitations, the intermediate components are difficult to be detected, therefore their detailed kinetics are still not well established. To address this issue, novel Chemistry-Informed Neural Networks (CINNs) were developed to derive the lignocellulosic biomass pyrolysis kinetics from the thermogravimetric analysis (TGA) measurements in published literature. The derived pyrolysis kinetics, involving eight species and eleven reactions, could accurately reproduce the pyrolysis process for both the seen and unseen samples with R2>0.95. The comparisons with the CRECK multi-step and Bio-CPD models also demonstrated the advantages of the derived kinetics in predicting both the final volatiles yield and the pyrolysis process for various biomass types. This study explored a new tool for establishing solid fuel conversion kinetics from TGA measurements using chemistry-informed machine learning approaches.

Effect of Schmidt number on mass transfer across a sheared gas-liquid interface in a wind-driven turbulence.

The mass transfer across a sheared gas-liquid interface strongly depends on the Schmidt number. Here we investigate the relationship between mass transfer coefficient on the liquid side, kL, and Schmidt number, Sc, in the wide range of 0.7 ≤ Sc ≤ 1000. We apply a three-dimensional semi direct numerical simulation (SEMI-DNS), in which the mass transfer is solved based on an approximated deconvolution model (ADM) scheme, to wind-driven turbulence with mass transfer across a sheared wind-driven wavy gas-liquid interface. In order to capture the deforming gas-liquid interface, an arbitrary Lagrangian-Eulerian (ALE) method is employed. Our results show that similar to the case for flat gas-liquid interfaces, kL for the wind-driven wavy gas-liquid interface is generally proportional to Sc-0.5, and can be roughly estimated by the surface divergence model. This trend is endorsed by the fact that the mass transfer across the gas-liquid interface is controlled mainly by streamwise vortices on the liquid side even for the wind-driven turbulence under the conditions of low wind velocities without wave breaking.