Seeking a clean alternative energy resource is inevitable because of the limited fossil fuel energy resources and greenhouse gas emissions issue. Recently, advances in chemical and fuel processing technologies allow us to convert biomass to energy products with high energy density and value. Fast pyrolysis process is among the promising technologies for converting biomass to bio-oil and combustible gases and has gained substantial attention due to its ability to produce high yields of bio-oil, a valuable liquid which can be further upgraded to transportation fuels. Nonetheless, many obstacles need to be overcome in order to utilize biomass fast pyrolysis effectively and economically. For example, moving to large-scale operations is an important step to lower the capital cost of such processes. However, a detailed understanding of the complex thermo-physical phenomena happening inside the fast pyrolysis reactors is needed for designing and optimizing the process at large scales.

In this work, biomass fast pyrolysis is studied in various reactor geometries using a comprehensive numerical framework developed in this study. In this framework, a combination of a flow solver and chemical reaction solver is employed to describe pyrolysis of biomass. A multi-fluid model is used to describe the multiphase hydrodynamics of fast pyrolysis and the kinetic theory of granular flows is used to account for the solid phases. Then, a global pyrolysis reaction mechanism is coupled with the multi-fluid model to build a comprehensive CFD model capable of predicting time-dependent properties of chemically reacting multi-phase flows in pyrolysis process. A time-splitting technique is also employed to couple the flow solver and reaction kinetics. This numerical model is first tested on a bubbling fluidized bed pyrolyzer and validated using experimental data from literature. Simulation results for pure cellulose and red oak pyrolysis in bubbling fluidized bed reactors show good level of agreement with experimental values. Moreover, zero-dimensional modeling of biomass fast pyrolysis is carried out by estimating the vapor residence time in the bubbling fluidized bed reactor simulated in this study. Later, a single-auger reactor is studied using the present CFD model and results are validated using experimental data obtained from the auger reactor experiment at Iowa State University. Finally, the effects of operating conditions on the product yields are investigated in a single-auger reactor. Operating variables including reactor temperature, nitrogen flow rate, biomass feed rate, biomass pre-treatment temperature, reactor length and reactor diameter are varied and their effects are characterized. Numerical results show that extremely high reactor temperatures (> 823 K) favor syngas formation and decrease tar and unreacted biomass yields. While increasing nitrogen flow rate and shorter reactor lengths produced favorable results. Similar to experimental data, numerical simulations also show that using thermally pre-treated biomass results in higher yields of syngas and lower unreacted biomass and tar yields. Simulations indicate that the auger reactor configuration is very sensitive to biomass feed rate, resulting in high yields of unreacted biomass when high biomass feed rates are applied. To address this issue, a single-auger reactor with larger diameter compared to the standard auger is simulated and resulted in substantially lower unreacted biomass yield.