Conference Dates

March 8-13, 2009


From escalating global concern over the exhaustion of non-renewable energy sources comes an imperative need for the use of renewable resources. Among possible renewable sources and subsequent conversion processes, biomass pyrolysis is a very promising alternative for fuel and chemical production. Previous work has shown that the most effective biomass pyrolysis processes employ downer reactors. Ongoing research at the Institute for Chemicals and Fuels from Alternative Resources (ICFAR) has led to the development of a new downer reactor design for the pyrolysis of biomass feedstock. However, this process requires a gas-solid separator that achieves a minimum spread of the gas residence time distribution – i.e. as near to plug flow as possible.

A novel integrated gas-solid inertial separator has been designed for implementation in the downer reactor. This new separator combines both primary separation and solids stripping within the same device. This is intended to decrease the product vapor residence time as well as to reduce the severity of vapor overcracking as compared to other separation methods proposed in the literature. The gas-solid separation section of the new device features a uniflow configuration and a vertical, axial entry with swirl vanes.

In the current study, various geometry configurations and operating conditions were tested for their effect on separation efficiency and pressure drop. A 5 cm-diameter separator was tested under cold flow modeling conditions using air, silica sand and glass beads. The sand and glass beads had Sauter mean diameters of 200 µm and 63 µm, respectively. The gas inlet velocity was varied from 1 to 21 m/s and the solid load ranged from 1.1 to 21 wt./wt. The separation length was adjusted from 0 to 2.5 separator diameters.

Initial cold flow experiments using silica sand revealed that the separator performance was influenced greatly by separator geometry. Flow deflector blades (i.e. swirl vanes) induced only a very weak swirl to the incoming flow. Hence the primary means of particle collection at the wall was by particle deflection as opposed to imparted centrifugal forces. The optimum separation length was determined to be zero (i.e. deflector blades positioned very near the gas outlet). Solids recovery in excess of 99.99% was achieved at zero separation length. The solids recovery decreased steadily as the separation length increased, to a minimum of 99.78% at 2.5 separator diameters. At long separation lengths, visual observation confirmed that particles were re-entrained in the exiting gas flow above the gas outlet after colliding with the separator wall. Several blade geometries were tested, where the most effective separation was obtained using 30° deflector blades (as measured from the separator vertical axis). Solid recovery was in general not strongly affected by gas inlet velocity or solid loading. Finally, the separator grade efficiency curve was measured from experiments using glass beads.