(423b) Reduced Order Particle-Scale Model for Biomass Pyrolysis and Gasification in Fluidized Bed Reactors

Bakshi, A. - Presenter, National Energy Technology Laboratory
Faravelli, T., Politecnico di Milano
Stark, A. K., US Department of Energy
Altantzis, C., National Energy Technology Laboratory
Ghoniem, A. F., Massachusetts Institute of Technology
Thermochemical conversion of biomass is a viable route for the production of renewable fuels. Current technologies face challenges efficiency and scaling challenges, due in part to the lack of accurate process and reactor models. Computational Fluid Dynamics (CFD) can enable comprehensive simulations for optimizing reactor designs. One of the biggest challenges with reactive simulations, however, is the coupling of particle-scale phenomena with reactor-scale processes. Typically, global kinetic mechanisms are employed which are (a) feedstock specific and cannot be generalized and (b) do not capture key transient dynamics such as the evolution of primary pyrolysis.

This work is focused on (a) the fundamental investigation of biomass particle conversion and (b) the development of a reduced order model to enable robust coupling with reactor-scale processes. Towards this end, high-fidelity particle scale CFD simulations are conducted using bioSMOKE, developed at Politecnico di Milano. BioSMOKE captures the evolution and thermal degradation of arbitrarily-shaped woody biomass particles based on detailed kinetic mechanisms. Operating conditions are chosen typical to fixed and fluidized bed reactors: particle diameter in the range 0.1-5.0 cm, pyrolysis temperatures 800-1200 K and heat transfer coefficient 100-500 W/m2-K. Using detailed statistics of the conversion phenomena, we show that the state of biomass and evolution of major volatile species (CO, CO2, H2, tar, etc.) can be parameterized using the mass fraction, Biot and pyrolysis numbers. This phase space is used to develop a reduced order model for biomass conversion which predicts the transient evolution of major species and char in excellent comparison with raw simulation data over the wide range of operating conditions investigated. The framework developed in this study allows for comprehensive characterization of particle scale phenomena within feasible computational time, without making any assumptions regarding the particle isotropy or reactivity whatsoever. The model can be easily incorporated into existing reactor-scale CFD simulations enabling robust simulations of fluidized and fixed bed reactors.