Analysis of Fluid and Particle Dynamics in a Spiral Jet Mill by Coupled CFD-DEM

The spiral jet mill is fundamentally very simple in its design and operation. However, it is difficult to predict the particle product size without empirical knowledge of the system being used. There is a dynamic relationship between solid material and gas inputs, as well as an interdependence between breakage and classification all occurring in the same chamber. The role of particle hold-up and how it affects both particle dynamics and fluid behaviour has been investigated. By changing the mass loading and keeping the grinding gas pressure constant it is shown that the particle bed indirectly decreases the fluid gas velocity surrounding the classifier, as the particles in the bed dampen the fluid field and lower its kinetic energy. It is also shown that the kinetic energy of the particle system remains largely unaffected by mass loading, whilst the vortex is stable. This is because the energy provided to the bed by the gas jets remains unchanged.
The grinding gas pressure is varied at different mass loading to investigate the jet behaviour. The results show that particle behaviour changes as the number of particles increases, or pressure decreases, as the grinding gas jets can no longer penetrate through the densely packed bed. This inhibits particles from reaching the highest velocities needed to cause fragmentation and chipping.
To approach scaling, the coarse grain method is applied. This method replaces groups of smaller particles, with a single larger one. Both the kinetic energy and dissipated energy are modelled successfully for the lowest scaling values of 4 or fewer particles per group. The predicted behaviour of the particles in the bed agrees with the base case. However, particle velocity in the lean region disagrees with the original simulation due to the reduced number of particles present. Finally, the Ghadiri & Zhang breakage model has been implemented to allow for size reduction by chipping. By recording the mass loss with respect to time at various gas pressures, the total work leading to breakage can be calculated and optimised. The work developed here provides a capability of predictive milling for various applications based on the mechanical and physical properties of single particles.