The thermochemical conversion of biomass to biofuels via fast pyrolysis has been of growing interest in recent decades. The conventional method is done in the absence of oxygen, and the pyrolysis of biomass is an endothermic process. The heat needed for the process is then supplied from the walls of the reactor. While supplying heat in this manner may be possible for small-scale reactors, it quickly becomes unfeasible for plant-scale reactors due to the reduced surface area to volume ratio of the reactor. Heat transfer becomes a bottleneck for the scale-up of the conventional fast pyrolysis of biomass^1,2.

Autothermal biomass pyrolysis processes¹ are being developed to address the bottleneck to biomass pyrolysis scale-up. Small amounts of oxygen are injected into the fluidized bed reactor, which allows for partial oxidation^1,3 of pyrolysis products. The exothermic reactions supply heat for the endothermic reactions, such as devolatilization, balancing the energy requirements for the process. One of the major contributors of heat in the fluidized bed is char combustion, and the retention of char in the bed is essential for its reaction with oxygen. Simultaneously, the quick release of volatile gases from the biomass is essential for bio-oil yield. The performance of autothermal pyrolyzers may be affected by several design choices of the fluidized bed reactor. For example, the height of the injection point of biomass may influence the mixing of biomass and sand in the reactor, affecting the residence time of the biomass and consequently the product yield.

To investigate these effects, we have developed a computational model to describe biomass fast pyrolysis. The model is based on the Eulerian multi-fluid model^4,5, which originates from a random field statistical approach for gas-solid glows. The phases are treated as interpenetrating continua that obey conservations of mass, momentum, and energy. The biomass and sand are granular phases, and the closure for solid granular properties such as pressure, viscosity, and energy are obtained from the kinetic theory of polydisperse granular flows^6,7 and frictional stress models^8,9. The model is implemented into OpenFOAM®. A comprehensive chemical kinetics mechanism^10–13 to account for devolatilization, char combustion, and secondary gas-phase reactions was incorporated into the solver.

Initially, the hydrodynamics and the chemical kinetic mechanism were studied independently. The multiphase solver was used to study the mixing of biomass and sand in the fluidized bed. Three-dimensional cold-flow simulations of a lab-scale fluidized bed reactor without reactions were done. Two designs with different biomass injection locations are considered to study the impact on mixing of the biomass injection height. A new type of mixing index based on the average volume fraction field was developed to quantify the mixing from the multiphase CFD simulations. The mixing indices developed are fields that provide the spatial representation of gas-solid mixing and solid-solid mixing in the fluidized bed. These new indices prove to be a powerful tool to identify regions of rich and lean mixing. Conventional mixing indices such as Lacey and Rose are extended to be used with the new mixing index and provide a single scalar measure of mixing for the entire fluidized bed.

Two chemical kinetic mechanisms were taken into consideration, namely, the Miller-Bellan¹⁰ and the Ranzi^11,12 mechanisms. The Ranzi model is realistic because it has the information of gas species produced from the devolatilization of biomass. It consists of two aspects, (a) the devolatilization reactions, which account for the volatile gases and char that are first obtained from biomass, and (b) the secondary gas-phase reactions, which comprise the CRECK model where the volatile gases further react. The CRECK secondary gas-phase model has 137 species and 4533 reactions. The Miller-Bellan mechanism is a simplified mechanism, where gas species are lumped into single species for non-condensable gases and bio-oil vapors. Both mechanisms were developed to study multi-component biomass feedstocks made up of cellulose, hemicellulose, and lignin. Both mechanisms follow a similar pattern where first, the virgin biomass forms an activated component, often considered a depolymerization step. Then volatile gases and char are produced from these activated biomass species. Char combustion reactions were added to these mechanisms to account for the main heat source in the autothermal pyrolysis process.

The chemical kinetic models were implemented in the chemistry solver available in OpenFOAM. The time evolution of the species and heat of reactions were studied. The Miller-Bellan mechanism was modified to match the performance of the Ranzi mechanism. Isothermal, zero-dimensional simulations of the chemistry were performed for various temperatures to study both mechanisms. Different biomass feedstocks were considered, but the focus was on red oak and corn stover.

The chemical kinetic models were incorporated into the multiphase solver with the fluidized bed reactor. The lab-scale reactor has a diameter of 3.81 cm and is 42.7 cm in height. Sand was filled to a height of 10.5 cm, and the bed is initially stationary at a fixed initial temperature. Biomass is injected into the bed at 300 K from the reactor side, while the fluidizing gas enters from underneath the bed. The fluidizing gas comprises a mixture of air and nitrogen injected with a superficial gas velocity at the inlet. Biomass is injected at a rate of 1 kg/h, while the fluidizing gas enters the domain at 20 SLPM. Unlike the chemistry solver, the CFD simulations were not isothermal, but varying initial thermal conditions of the bed are considered as sensible heat is needed to raise the temperature of biomass form 300K to the desired pyrolysis temperature. The yield of bio-oil is computed for these conditions and contrasted with the conventional fast pyrolysis process.

The Ranzi mechanism, being an elaborate chemical kinetic model coupled with the multiphase CFD solver, is computationally expensive. The feasibility of such a comprehensive model as a predictive tool for the yields of bio-oils in the pyrolysis process comes under scrutiny. A plug flow reactor (PFR) model of the lab-scale pyrolyzer was used and showed little to no secondary gas-phase reactions with the CRECK mechanism for the size and operating conditions of the reactor. However, the secondary gas-phase reaction rates are low and depend on the residence time of the gas species in the fluidized bed reactor, which in turn depends on the solid-gas mixing. Therefore, based on the Damköhler number, which is the ratio of mixing to chemical reaction time scales, the secondary gas-phase reactions may be present in the reactor and captured by the CFD model. If these reactions are prevalent in the reactor, they should not be excluded from the model, as the heat from the reactions of the gas species will contribute to the autothermal pyrolysis process.

References

1. Polin JP, Peterson CA, Whitmer LE, Smith RG, Brown RC. Process intensification of biomass fast pyrolysis through autothermal operation of a fluidized bed reactor. Appl Energy. 2019;249:276-285. doi:10.1016/j.apenergy.2019.04.154
2. Proano‐Aviles J, Lindstrom JK, Johnston PA, Brown RC. Heat and Mass Transfer Effects in a Furnace-Based Micropyrolyzer. Energy Technol. 2017;5(1):189-195. doi:10.1002/ente.201600279
3. Kim KH, Bai X, Rover M, Brown RC. The effect of low-concentration oxygen in sweep gas during pyrolysis of red oak using a fluidized bed reactor. Fuel. 2014;124:49-56. doi:10.1016/j.fuel.2014.01.086
4. Drew DA. Averaged equations for two-phase flows. Stud Appl Math. 1971;L(3):205-231.
5. Drew DA. Continuum Modeling of Two-Phase Flows. In: Meyer R, ed. Theory of Dispersed Multiphase Flow. Academic Press; 1983:173-190.
6. Jenkins JT, Savage SB. A theory of the rapid flow of identical, smooth, nearly elastic spherical particles. J Fluid Mech. 1983;(130):187-202.
7. Jenkins JT, Mancini F. Balance Laws and Constitutive Relations for Plane Flows of a Dense, Binary Mixture of Smooth, Nearly Elastic, Circular Disks. J Appl Mech. 1987;54(1):27-34. doi:10.1115/1.3172990
8. Johnson PC, Jackson R. Frictional–collisional constitutive relations for granular materials, with application to plane shearing. J Fluid Mech. 1987;176:67-93. doi:10.1017/S0022112087000570
9. Srivastava A, Sundaresan S. Analysis of a frictional–kinetic model for gas–particle flow. Powder Technol. 2003;129(1):72-85. doi:10.1016/S0032-5910(02)00132-8
10. Miller RS, Bellan J. A generalized biomass pyrolysis model based on superimposed cellulose, hemicellulose and lignin kinetics. Combust Sci Technol. 1997;126(1-6):97-137. doi:10.1080/00102209708935670
11. Ranzi E, Debiagi PEA, Frassoldati A. Mathematical Modeling of Fast Biomass Pyrolysis and Bio-Oil Formation. Note I: Kinetic Mechanism of Biomass Pyrolysis. ACS Sustain Chem Eng. 2017;5(4):2867-2881. doi:10.1021/acssuschemeng.6b03096
12. Ranzi E, Debiagi PEA, Frassoldati A. Mathematical Modeling of Fast Biomass Pyrolysis and Bio-Oil Formation. Note II: Secondary Gas-Phase Reactions and Bio-Oil Formation. ACS Sustain Chem Eng. 2017;5(4):2882-2896. doi:10.1021/acssuschemeng.6b03098
13. Calonaci M, Grana R, Barker Hemings E, Bozzano G, Dente M, Ranzi E. Comprehensive Kinetic Modeling Study of Bio-oil Formation from Fast Pyrolysis of Biomass. Energy Fuels. 2010;24(10):5727-5734. doi:10.1021/ef1008902