EDEM-Fluent coupling

Fatemeh Hosseini_22198

Hello everyone.
I am studying particles deposition and agglomeration in a turbulent flow tube. I used k-omega method  as a turbulence model. 
The particles diameter is 50 micron.
The minimum volume of mesh is= 1.18e-11
particle time step=20% *Reyliegh=3.8e-8

Fluid time step=e-6
The boundary condition is periodic in both EDEM and Fluent too.
When I start simulation it diverges. because it can not get any information of particle from EDEM. (The screen shot has been attached)
Could you please guide me?

Thanks a lot

 

Stephen Cole

Hi Fatemeh,

 

Do both EDEM and Fluent run un-coupled? It is always best to confirm that both programs run OK prior to the coupling, if you see errors in one or the other it helps identify any issues.

 

Is the mesh volume above in meters?  If so that is OK, just to check that the mesh volume is greater than the particle volume.

Another area to check is that the geometry is the same scale in both programs, can you confirm that the dimensions/scaling for both are the same (you can check in EDEM via the Environment tab in the Creator).

 

Other than that it could be the time-step ratio, typical recommendations are 10:1 or 100:1 (with fluent having the larger timestep), at 100-1 this means that EDEM is solving 100 timesteps and it really depends how far the particle moves in this time frame, will it be jumping between mesh cells which could cause instability or would it transition between cells in a few Fluent times-steps.  If it is moving quickly between cells then a lower timestep ratio (lower timestep in Fluent) would be beneficial.

Regards

Stephen

Fatemeh Hosseini_22198

Stephen Cole_21117 said:

Hi Fatemeh,

 

Do both EDEM and Fluent run un-coupled? It is always best to confirm that both programs run OK prior to the coupling, if you see errors in one or the other it helps identify any issues.

 

Is the mesh volume above in meters?  If so that is OK, just to check that the mesh volume is greater than the particle volume.

Another area to check is that the geometry is the same scale in both programs, can you confirm that the dimensions/scaling for both are the same (you can check in EDEM via the Environment tab in the Creator).

 

Other than that it could be the time-step ratio, typical recommendations are 10:1 or 100:1 (with fluent having the larger timestep), at 100-1 this means that EDEM is solving 100 timesteps and it really depends how far the particle moves in this time frame, will it be jumping between mesh cells which could cause instability or would it transition between cells in a few Fluent times-steps.  If it is moving quickly between cells then a lower timestep ratio (lower timestep in Fluent) would be beneficial.

Regards

Stephen

Thanks for your answer Stephan.
Both of the program run well individually. 
I copy the scale in Fluent on the domain in environment in EDEM. So I think geometry is same. 
Changing time step doesn't have positive effect unfortunately. 

Best,
Fatemeh

Jani-Petteri Jylhä

Hi,

I've sometimes had similar issues even when models themselves seem to be great and run perfectly without coupling. One trick that may help at those times is to start simulating with just edem and when the first particles have been created, stop and start simulating with the coupling. Also, sometimes running fluent with a bit lower time step has helped and when the flow field is relatively stable, it seems safe to increase time step.

This either helps or it doesn't but I think it's worth trying.

 

BR,

Jani