fluidization in 90-degree bend
 

I am trying to simulate the pneumatic transport of solids in 90-Degree pipe bend. The mass fraction of the solids is about 0.99 and the including air about 0.01. We are interested in the simulation of the sedimentation process in dependece of the bulk mass flow at the outlet. Because of the high volumetric fraction of the solid phase (0.63) we need to use a euler-euler description. to simulate the sedimentation we need to switch to the inhomogeneous model with the Gidaspow drag law. If we include gravity and especially the "kinetic model" for the dispersed phase we get high instability while solving. At the moment we are thinking about an alternative approach to modell the sedimentation process. Does anybody have an idea about an alternative?

CFX has no built in sedimentation or scour model. So for very high particle loadings like this I suspect CFX is not going to be able to do it. CFX has a fluidised bed model which can deal with particle densities approaching this, but that never sedimentates or scours. It is designed for coal combustors and agitated sand beds and things like that.

Thanks for your comments.
There are some papers where CFX or Fluent is sucessfully used for simulating pneumatic transport. Currently we developed the simulation step by step. We got results of an inhomogeneous Euler/Euler simulation using Gidaspow Drag law and no particle collison model. The result is a solid layer with volume fraction of 1 that moves on the bottom of the pipe while the air moves with highe velocity. Next step was to implement the complete kinetic theory. Even with 1µs time steps the simulation crashes. So we switched off again and implemented the Gidaspow modell for solid pressure to limit the maximumung packing to 0.63. But even if we use the solid pressure model the volume fraction of the particle phase does not reach a maximum fraction of 0.63. Does anybody has an idea what could be the reason for that. Or does anybody have an idea about an more efficient approach as the kinetic theory to get the main flow behaviour?