[bonimio]

Hello. I'm trying to simulate a 3D fludized bed using a eulerian-eulerian approach. When I restrict the inlet area, fluent diverges and show me the message: floating point exception. I'm using gidaspow for granular viscosity and drag coefficient.

inlet velocity: 0.12 m/s
diameter of particles: 0.000163m
density: 2820 kg/m3
base radius: 0.073m
inlet radius: 0.01m
step size: 0.001
element size: 0.005

If my inlet area is the base area (0.073m of radius) the is no problem, but when I put the named section with 0.01m of radius, the problem diverge after 20 interections.

Thank you very much in advance

[Indra]

Originally Posted by bonimio

Hello. I'm trying to simulate a 3D fludized bed using a eulerian-eulerian approach. When I restrict the inlet area, fluent diverges and show me the message: floating point exception. I'm using gidaspow for granular viscosity and drag coefficient.

inlet velocity: 0.12 m/s
diameter of particles: 0.000163m
density: 2820 kg/m3
base radius: 0.073m
inlet radius: 0.01m
step size: 0.001
element size: 0.005

If my inlet area is the base area (0.073m of radius) the is no problem, but when I put the named section with 0.01m of radius, the problem diverge after 20 interactions.

Thank you very much in advance

I guess one of the immediate things to do is reduce time step size. What sort of pressure discretisation are you using? I would recommend using PRESTO! or Body force weighted. Also, you might wanna try switching to explicit scheme under multiphase model if you are running a trasnient sim. But don't change everything at once!

A general advice, in the tui you can type /report/mphase-summary and select 1 for verbosity. It will give you all sorts of recommendation and suumary

[bonimio]

Originally Posted by Indra

I guess one of the immediate things to do is reduce time step size. What sort of pressure discretisation are you using? I would recommend using PRESTO! or Body force weighted. Also, you might wanna try switching to explicit scheme under multiphase model if you are running a trasnient sim. But don't change everything at once!

A general advice, in the tui you can type /report/mphase-summary and select 1 for verbosity. It will give you all sorts of recommendation and suumary

Thanks for replying me . Since I saw your answer, I'm trying to simulate my case. I'm using PRESTO and change time step to 1^-5, but the simulation time incresed a lot, so I've change the element size to 1^-2. With those changes, I can simulate, but the residuals periodicaly show-me that (see the figure). Do you think changing the element size and reducing the time step, the behaviour will change?