Numerical Study and Experimental Validation of Particle Strand Formation
Sprache des Titels:
6th International Conference on CFD in Oil & Gas, Metallurgical and Process Industries
Pneumatic conveying of spherical glass particles through a rectangular channel is studied by means of numerical simulations and compared with optical measurements taken from the experimental setup. Thereby, a double – looping is placed in front of the straight channel in order to generate a particle strand at the bottom of the channel. Finally the profiles of particle velocity and volume concentration are measured by particle image velocimetry (PIV).
The corresponding unsteady, three – dimensional numerical simulations are carried out with the FLUENTÒ software package. Discrete particle paths are traced in Lagrangian reference frame. Nevertheless, some further sub – models have been introduced in order to describe different mechanisms of particulate flows in some more detail. These mechanisms include:
· Particle – wall collisions,
· Particle – particle collisions and
· The influence of particle rotation on their trajectories.
By passing the double – looping, the particles collide very often with the wall. Thereby, large rotational velocities are induced. Driven by the forces resulting from these rotational velocities the particles tend to move to the upper part of the channel right after the double – looping, thus dispersing the particle strand. This effect can be predicted by applying the particle – wall collision model and particle rotation during the simulations. In a second step, particle rotation is damped through particle – particle collisions, so the particle movement to the top of the channel is reduced. Furthermore, interparticle collisions are enhanced through high particle rotational velocities, because the velocity component perpendicular to the main stream direction is increased.
The full application of the particle – particle collision model is rather time consuming, especially in case of high mass loadings. Nevertheless, these sub – models are crucial for the proper prediction of the particle strand formation.