What caused the removal of the majority of the particles from the domain in EDEM-Fluent coupling?
Hello, everybody,
I am trying to simulate particles in a turbulent flow by Coupling Fluent and EDEM.
Both Fluent and EDEM run well individually.
Initially, after over 1000 iterations of running the CFD-DEM simulation, the EDEM crashed.
The mesh in Fluent had been verified and the CFD/DEM timestep ratio was 80.
In order to prevent reverse flow at first, I tried to shift the particle's entrance and decreased the CFD timestep (CFDtimestep/DEM timestep=50).
It appears to be running now, but most of the particles were removed.
The boundary condition is periodic and all the particles should remain on the tube. (Please see the attached picture)
Additionally, the maximum velocity for particles is 190 m/s, but the maximum velocity for fluid particles is 15 m/s!
Would you kindly advise me on this matter?
Regards,
Fatemeh
Best Answer
-
Fatemeh Hosseini_22198 said:
Dear Stephane,
Thanks for your response.
I checked the particle distance in one Fluid time step. It is less than the mesh size.
The time-step ratio is now 32 after I again reduce the CFD time step!
The simulation crashed again after 12300 iterations. I can't follow the particle in EDEM because it doesn't save the data!!
This is the velocity diagram in Fluent when it crashed. (Please see the attached picture)
Do you think the time-step ratio you mentioned earlier played a role in what happened?
(I think 32 is small enough for the timestep ratio)
Regards,
Fatemeh
Hi Fatemeh,
It still suggests it maybe time-step related instability, you could try a 1:1 time-step ratio just to check stability.
You can change the save interval period in EDEM to save more data to help with investigating the model, you don't want to save every time-step but if you can save enough steps within the 0-0.02 s time-period to track the particles this should give you a good idea of what is happening in the system.
RegardsStephen
0
Altair EmployeeAnswers
-
It crashed after 3709 iterations again!
0 -
Hi Fatema,
The high velocities are still most likely to be time-step or time-step ratio related. What you often see if the ratio is too high is the particles will accelerate beyond the fluid speed, then try to decelerate too much the next fluent time-step ad then accelerate further again until the simulation becomes unsteady.
It is often best to check where/when the issue is first happening by reviewing the EDEM playback step by step to see if you can identify this behavour.
Also another check you can make is an estimation of how far the particle will travel in 1 fluent time-step (2e-6 s in this case?) at an estimation of the maximum speed expected (15 m/s?). In a dynamic system really you want this distance to be smaller than the mesh cell length so that the particles will exist within a cell for more than 1 fluent time-step.
RegardsStephen
1 -
Stephen Cole_21117 said:
Hi Fatema,
The high velocities are still most likely to be time-step or time-step ratio related. What you often see if the ratio is too high is the particles will accelerate beyond the fluid speed, then try to decelerate too much the next fluent time-step ad then accelerate further again until the simulation becomes unsteady.
It is often best to check where/when the issue is first happening by reviewing the EDEM playback step by step to see if you can identify this behavour.
Also another check you can make is an estimation of how far the particle will travel in 1 fluent time-step (2e-6 s in this case?) at an estimation of the maximum speed expected (15 m/s?). In a dynamic system really you want this distance to be smaller than the mesh cell length so that the particles will exist within a cell for more than 1 fluent time-step.
RegardsStephen
Dear Stephane,
Thanks for your response.
I checked the particle distance in one Fluid time step. It is less than the mesh size.
The time-step ratio is now 32 after I again reduce the CFD time step!
The simulation crashed again after 12300 iterations. I can't follow the particle in EDEM because it doesn't save the data!!
This is the velocity diagram in Fluent when it crashed. (Please see the attached picture)
Do you think the time-step ratio you mentioned earlier played a role in what happened?
(I think 32 is small enough for the timestep ratio)
Regards,
Fatemeh0 -
Fatemeh Hosseini_22198 said:
Dear Stephane,
Thanks for your response.
I checked the particle distance in one Fluid time step. It is less than the mesh size.
The time-step ratio is now 32 after I again reduce the CFD time step!
The simulation crashed again after 12300 iterations. I can't follow the particle in EDEM because it doesn't save the data!!
This is the velocity diagram in Fluent when it crashed. (Please see the attached picture)
Do you think the time-step ratio you mentioned earlier played a role in what happened?
(I think 32 is small enough for the timestep ratio)
Regards,
FatemehHi Fatemeh,
It still suggests it maybe time-step related instability, you could try a 1:1 time-step ratio just to check stability.
You can change the save interval period in EDEM to save more data to help with investigating the model, you don't want to save every time-step but if you can save enough steps within the 0-0.02 s time-period to track the particles this should give you a good idea of what is happening in the system.
RegardsStephen
0