[CeesH]

Hello all,

I'm working on particle tracking and keep getting the error:

"Cannot find intersecting face for particle - ABORT..."

Now, I've one quite a few DPM simulations already and have never seen this error. I've been trying to get rid of it by redefining interfaces and so on, but to no avail. Does anyone have an idea what causes this errors, and better, a solution?

Cheers
C

[`e`]

I've not seen this error either, have you modified your mesh in Fluent and where did you create the mesh? I'm guessing the error is related to the mesh. Is the continuous phase being solved without problems? Lastly, try running Fluent with the serial solver.

[CeesH]

hi e,

I had modified my mesh in FLUENT indeed, so I went back to the original... and found the same problem. Also, I could identify the faces where it happened, they were nowhere near the modified section. (the faces were that of the impeller, which is modelled as a thin sheet wall)

Continuous is solved fine, error happens also in serial.
In the end, I just ran. The only thing I see is that due to the abortion the residence time of some particles is shorter than of others, but other than that, the trajectories look realistic, and the small changes in time don't hurt my results. But I'm still curious as to what happened, since my meshing and solving procedure were equal to earlier runs...

[`e`]

Sounds like it's an issue when calling the boundary condition for the DPM. Perhaps try using escape, trap, reflect etc as the boundary conditions for the DPM on the impeller (if that's an option with thin sheet walls). If you're only getting this error now and it was fine before then there must have been something that changed (albeit maybe not obvious).

[CeesH]

It is reflect by default, but during particle tracking I change the coupled walls of the impeller to a single wall and set it to interior (otherwise the particles stick to the impeller due to vanishing U, k and e at the wall - that thing works better than fly paper in capturing small moving objects). But also this is my default approach, so yeah, still trying to figure out what changed.. (the baffles are treated in a similar way, and there no error pops up)

[CeesH]

Got it; the simualtion I ran was a gas-liquid phase run with particles in the liquid phase. So, particles passing through the blade would jump into the trailing vortex, where the liquid volume fraction is 0, and with the particles confined to the liquid phase, they cannot make this jump.

[CeesH]

Ok, I'm sort of retracting the previous post. Problem keeps popping up, even when I set the impellers back to walls - so the origin is not particles popping through...