Run Divergence By Velocity, STL or Particle Seeding Error?

MFiX
#1

I am modeling a simple thin solid surface for 10 particles to transfer heat onto. The goal is to obtain the temperatures from the surface as a result of contact with the higher temperature particles.

The model has an error within the first minute of the run due to DT<DT_MIN.

I received a few different warnings prior to the stop, but I believe it stems from the particles not being seeded

Warning 1202 The system is initiated with no particles and no solids inlet was detected.

and out of bound velocities. I have tested higher scale factors, but that also scales the velocities up to meet it.

Warning: velocity higher than maximum allowed velocity: 50.000 (to change this adjust the scale factor MAX_INLET_VEL_FAC)

Finally, the run stops with this error. I saw from previous forum posts that this may be a mesh error, however the imported stl is a prismatic plate, basically a box so I don’t image there is much complex meshing that would be generating errors.

Error from time_step.f:193
DT < DT_MIN. Recovery not possible!

This is the first time I have posted on this forum, so if there is any additional information needed, please let me know what I can provide. Thank you!
100p_surfacemodeled.mfx (21.0 KB)
100P_SURFACEMODELED.msh (9.3 MB)

#2

Please include the stl file. A easy way to bundle all necessary files is to go to the main menu and select “Submit bug report” (even though this may not be a bug). This will generate a .zip file that can be attached to a post.

A common error when using STL files is to have the normals pointing in the wrong direction. This could lead to not being able to seed particles is there are no fluid cells.

image

#3

Thanks for the tip @jeff.dietiker , the file has been attached below. I have also tried flipping the normals in the Geometry tab and the current configuration is the one that allows the file to begin the run. Checking the box creates a an error with no standard fluid cells.

100p_surfacemodeled_2022-01-17T084524.720254.zip (118.4 KB)

#4

A few observations:

  1. Your particle size range is too wide. With a mean diameter of 16mm, you should not go as low as 1 mm for the minimum size. Staying with 3 std dev away from the mean is typically sufficient to represent 99.7% of the distribution (see shaded area in the PDF plot). Here if you want to manually set the min and max, you could go say from 10 to 20 mm, which would give a ratio max/min size of 2. Setting a min of 1 mm will drive the time step down unnecessarily.

  2. The stl file is a rectangle align with major axes so you do not need it and it doesn’t help define the geometry. You should just use the regular box geometry.

  3. This is linked to 1) The des grid is set to 1x1x1 cell, probably because you used a large max diameter. This is the root of the failure to seed particles as it triggers false detection of stl facets in all des grid cells (there is only one cell, and all stl facets are attached to it). I can seed 10 particles with a des grid of 10x5x10 and your current pdf. My suggestions to address 1) and 2) first though.

image
image639Ă—536 85.6 KB

1 Like
#5

Thank you Jeff. Implementing 1 and 2 fixed the error.