(432b) Modeling the Impact of Biomass Particle Residence Time on Fast Pyrolysis Yield and Composition

Bubbling fluidized bed and circulating fluidized bed reactors are the most common unit operations for carrying out fast pyrolysis of biomass. These reactors are designed to contact biomass particles with hot inert gas (such as nitrogen) in a bed of solids (sand) under turbulent multiphase flow conditions. In bubbling bed reactors, the gas flow is sufficient enough to cause mixing of the solids and gas without entraining most of the particles, except for char and ash generated from the biomass which can elutriate out of the sand bed then exit from the top of the reactor. In circulating bed reactors, the gas flow is much higher and the solids are typically smaller, consequently most of the solids are entrained out the top of the reactor vessel. The entrained solids are separated from the gas and most of them are recirculated back through the reactor with new gas, hence the name circulating fluidized beds.

Numerous studies have demonstrated that biomass fast pyrolysis yields depend on the residence time distribution (RTD) of the pyrolyzing particles, which is a function of the inherent feed properties as well as the reactor design and operating conditions. Experimental measurements indicate that it should be possible to represent the RTDs for biomass particles in both bubbling and circulating bed pyrolysis reactors with relatively simple functional approximations involving only two or three physically relevant parameters. We expect similar approaches can also be used to model the impact of catalyst particle residence time in vapor phase upgrading reactors.

In this presentation, we summarize literature on low-order reactor models for the particle RTDs in bubbling and circulating fluidized bed reactors and discuss the implications for the conversion of biomass particles undergoing fast pyrolysis. Based on this information, we demonstrate how particle RTDs can be combined with reaction kinetics and simplified reactor mass balances to create low-order fast pyrolysis models that can be used to summarize, interpret, and extrapolate observed pyrolysis trends in experiments and more detailed computational simulations. Our goal is to provide flexible analytical tools for consolidating information from more detailed computational reactor simulations and improving high-level process simulations of bio-oil production needed for techno-economic analyses and reactor design.