Floating-point exception - erroneous arithmetic operation for DEM modelling

[sayyamchopra]

Hello @cgw ,

I am trying to run a DEM particle flow simulation and I have been facing a floating point exception issue. I am unable to understand what value should be put for (des_epg_clip). I have tried almost all the values from 0-1. Can you clarify a few of my concerns?

1. I see this is a new parameter put in the recent update because I have been using MFIX for 2 years now. The system never asked for this value. Can you please explain what this value is?
2. Now, how can I solve this issue?
3. Is the floating point exception due to this value of something else ? My simulation was running great until this new version 22.4.2 was released.

simulation3.5mm25secvar_velocity.mfx (17.3 KB)
geometry.stl (719.6 KB)
geometry_0001.stl (8.7 KB)
geometry.stl (719.6 KB)
chemicalloopfinal.stl (431.7 KB)

BOX1_FACETS_IGNORED.stl (25 Bytes)
SIMULATION3.5MM25SECVAR_VELOCITY.msh (27.0 MB)

[jeff.dietiker]

For DEM, we used to use EP_STAR to clip the void fraction. Starting in 22.4, we now use DES_EPG_CLIP (EP_STAR is now only used with TFM). This is used to avoid unphysical values of the void fraction, primarily is small cut cells, where, the void fraction could become unphysical. I think this is what is happening in your case. You did not define DES_EPG_CLIP, and you get an unphysical void fraction value in a cut cell. You can use a value similar to what is used with TFM, say around DES_EPG_CLIP=0.42.

(Please note that your mesh is too coarse to resolve the cyclone vortex finder.)

[sayyamchopra]

Thank you @jeff.dietiker for clarifying everything.

Sayyam