EDEM-Fluent coupling problem.

bai

Hi Everyone,

I have been trying to implement the Bonded-Particle contact model (v3.5.0) API provided by EDEM (from its previous EDEMSimulation forums), then it works fine. At the same time, I also tried to establish a coupling interface between edem2023 and fluent2022r1, and succeeded. When I tried to use the bead chain model created using the (v3.5.0)API and the simple flow field established by fluent for two-way coupling, the residual error was too large, as shown in the figure below.In contrast, when I simply use the built-in Bonding V2 model against the same parameters as for API, then it works fine.

Any suggestions, what could be the potential issue with that?

Thank you.

Best regards,

Bai

Stephen Cole

Hi,

The physics model used, either API or in-built, shouldn't directly influence the coupling as all Fluent can 'see' is the volume fraction and velocity/momentum of the particles.

IF you are changing physics models and you are seeing problems it is likely either due to large volume fractions in small fluent mesh cells or due to an unstable simulation where the particle velocities are high which would cause instabilities.

So I would check the EDEM alone simulation is stable, you can reduce the time-step used in EDEM and Fluent but often reducing the time-step ratio helps with stability. It's typical to have the Fluent time-step 10 or 100 times larger than EDEM but this means the forces are applied for this 100 time-steps before been updated, reducing this ratio means more CFD time-steps but increased stability.

Regards

Stephen

bai

BondedParticleCustom_3_5_0.zip

edem_coupling_build.zip

text_1_api_v3-5-5.zip

text_2_bonding_v2.zip

The attachment is as follows.

Stephen Cole

Hi,

The physics model used, either API or in-built, shouldn't directly influence the coupling as all Fluent can 'see' is the volume fraction and velocity/momentum of the particles.

IF you are changing physics models and you are seeing problems it is likely either due to large volume fractions in small fluent mesh cells or due to an unstable simulation where the particle velocities are high which would cause instabilities.

So I would check the EDEM alone simulation is stable, you can reduce the time-step used in EDEM and Fluent but often reducing the time-step ratio helps with stability. It's typical to have the Fluent time-step 10 or 100 times larger than EDEM but this means the forces are applied for this 100 time-steps before been updated, reducing this ratio means more CFD time-steps but increased stability.

Regards

Stephen

bai

https://community.altair.com/discussion/comment/188108#Comment_188108

Hi Stephen,

Thank you very much for your valuable advice. I ran edem alone and it was stable. Reduce the time step of fluent according to your suggestion. Only when it is reduced to 1e-9s can it run stably, for those using API. However, using bonding v2 can achieve a stable state when the time step is 1e-4s. This gap is relatively large, is there any good suggestion to improve this problem?

Best regards,

Bai

Stephen Cole

Hi Bai,

1e-9 s time-step does seem low, what is the time-step used for the EDEM model? It should only be reduced to lowest value = edem time-step.

In general for fluent stability reducing the under-relaxation factors, reducing these also helps with stability if it's related to convergence and mass/momentum transfer but best to check the Fluent documentation on impact of that.

Regards

Stephen

satri

Hello Bai,

I saw that you wrote this function called "ADAPTOR_registerCustomProperty" in this, what are the names of the custom properties that you wrote?

I want to bring this custom property value from fluent to EDEM which i want to change in EDEM. So i want to make sure both have the same name in EDEM and fluent.

I dont see anywhere you gave any name to the custom property.

You are registering the custom property in fluent at line 487. what are you naming it? and where are you using this custom property inside Fluent?

you are adding custom properties here, but how are you manipulating it? Where are the names of the custom properties that you gave?

Suppose you want to bring this custom property to fluent how will you bring it.

Can you please explain that to me. I cannot find any reliable resources, and I am not getting any good answers that i can understand and write the udf from the Altair discussion forum.

bai

https://community.altair.com/discussion/comment/189231#Comment_189231

Hello satri,

I apologize for not being able to answer your question, as I am just a beginner. Let me provide you with some background information about the case study where I posed the question, in hopes that it might be helpful.Based on the API (v3.5.0) provided by the EDEM staff, I developed a simple fiber model to replace Bonding V2, and then coupled it with Fluent to simulate a simple pipe flow. The approach to using the API involved compiling it into a DLL file using Visual Studio (VS), which the EDEM software could then call.

Best regards,

Bai

satri

Hey I understand that part where you compile in VS and use the bond model

Did you use the GUI provided by EDEM to compile the udfs? Did you modify modify the ansys udfs?

I am able to bypass the coupling GUI now and compile udfs directly with the Windows_coupling.py file. Only problem I have right now is there's some environment variable of ansys that is giving some kind of Cortex error in nmake did you encounter any such problem?

If so how did you fix it