[xeusater]

Hi everyone;

I am having a problem with an eulerian multiphase simulation. I am fluidizing a granular solid in a column reactor.
The 2D domain is a rectangle with the inlet at the bottom, outlet at the top and two free slip walls in the flanks. The granular phase is modeled basically using syamlal-obrien models for almost every parameter.

Solid particle density is about 1500kg/m3 and the gas is approx 4kg/m3 (heavy). The gas velocity is 0.212m/s at the inlet, which should provide an bubbling bed.

Now what happens is that as soon as I start the calculation, I get a huge backflow from the outlet. That is immediately, first iteration, on all faces of the outlet. Eventually the timestep converges, but the gas velocity field is facing downwards, at something like 400m/s.

Now, I have tried changing the timestep up to very small values (1e-9) , changed every Underrelaxation factor to many possible values, tried initializing the gas velocity field at 0m/s or 0.212m/s. Nothing works.

If there is no solids in the domain, then its Ok. But as soon as there is, even if the volume fraction equation is not solved, the problem starts.

Can you help me in this issue?
Thank you very much in advance