A reactive
computational fluid dynamic-discrete element model (CFD-DEM)
is coupled with a one-dimensional thermally thick model to simulate
biomass pyrolysis in a three-dimensional rotary drum reactor. Furthermore,
a diffusion-based voidage algorithm and a competitive pyrolysis scheme
are also implemented. After validation, the integrated model is applied
to explore the effects of the drum rotating speed and the filling
level. The particle distribution and motion, heat transfer, and conversion
behavior are selected as indicators. Results show that the particle
inventory within the central transverse section is affected by both
the mechanic effect of rotation and the axial conveying due to the
product gas flux. The angle of repose for the central section first
increases and then decreases with rotating speed, while increasing
the filling level reduces the angle of repose and delays the bed sinking.
Moreover, the particles at the bed surface are energetic and thus
the area of high-value granular temperature is distributed along the
bed surface. The particle accumulated displacement increases and the
particle residence time at the wall reduces with the rotating speed
and the filling level. In addition, the specific heat conduction rate
can be promoted by decreasing the rotating speed and the filling level.
A higher rotating speed or a lower filling level is favorable for
particle mixing and thus makes particles heat up uniformly. Finally,
a higher rotating speed or filling level exerts a negative impact
on biomass conversion. These findings are helpful for better understanding
biomass pyrolysis in a rotary drum reactor.