Expanded graphite particle is characterized by the low density in
comparison with those of bead glass and copper particles. Hydrodynamics
of the irregular-shaped graphite particle swirling flows in a coaxial
chamber are investigated via an improved kinetic frictional stress
model. A drag force coefficient considering the effects of irregular
shapes based on the artificial neural network algorithm is adopted
to describe the momentum transfer between nonspherical particles and
gas phases. The proposed model, algorithm, and source code for modeling
and simulation are validated by measurement using spherical glass
beads, and acceptable agreement is obtained. Lower sphericity particles
enhance the anisotropic particle dispersions and induces the redistributions
of the Reynolds stresses of the two-phase flow. Irregular-shaped particles
are more sensitive to the gas followability instead of own inertia,
whereas spherical particles are easier to be affected by the inlet
effects. The interlock force between nonspherical particles takes
great effect on particle flow than the spherical particle. The axial–axial
normal stresses of sphericities of 0.63 and 0.72 are approximately
3.4 times larger than those of shear stress of spherical particle,
and their axial velocities locating at near central regions are 3.0
times larger than those of sphericities.