MFiX 22.2.1 is available for download

[cgw]

MFiX version 22.2.1 is available for download. This is a minor update which fixes a few small issues from the 22.2 release.

1. Default solver not found on Linux
2. Fortran compiler not found on OSX
3. Solver build error scanner not working on Windows
4. Indexing error in PIC solver BC code

All users are encouraged to update, especially if you are affected by one of these issues.

[wuming]

Hi, Charles

Will this PIC solver update help to solve this problem?

Run diverged when simulating particle segregation using PIC method

Hi, I am simulating the segregating process of two particle phases having large density difference. What is interesting i found is that by changing the sizes of particles, the simulations diverged at different flow times, which may be due to the segregation rate of different particle system. What’s more, when particles are not segregated, the simulation ran well. I am not sure that this issue is resulted from my settings or the possible bug in MFIX-PIC code. Can you please check it? geometry.s…

[cgw]

@wuming - For reference, this is the PIC solver bug that was corrected between releases 22.2 and 22.2.1

commit 4f51fe7d3c64b6d153e87913757ff54e14dccf98
Author: Jeff Dietiker <jean.dietiker@netl.doe.gov>
Date:   Fri Jul 29 14:33:47 2022 -0400

    Fix bug (use des grid indices, not fluid indices).

diff --git a/model/des/pic/apply_wall_bc_pic.f90 b/model/des/pic/apply_wall_bc_pic.f90
index 27a25cfa9..dfb55777b 100644
--- a/model/des/pic/apply_wall_bc_pic.f90
+++ b/model/des/pic/apply_wall_bc_pic.f90
@@ -199,7 +199,7 @@ SUBROUTINE APPLY_WALL_BC_PIC
         dg_j = min(dg_jend2, max(dg_jstart2, jofpos(des_pos_new(np,2))))
         dg_k = min(dg_kend2, max(dg_kstart2, kofpos(des_pos_new(np,3))))
 
-        dg_ijk = dg_funijk(i,j,k)
+        dg_ijk = dg_funijk(dg_i, dg_j, dg_k)
 
         ! Loop through the STLs associated with the DES grid cell.
         collisions = 0

I do not know if this is related to the divergence you reported back in July. I tried running that case with 22.2.1 and it has not yet diverged. However I have only reached t=0.4s and I see that your original divergence happened at around t=2s. I’ll leave it running and see what happens. I suggest you also try it for yourself.

– Charles

[cgw]

It’s gotten to t=3s without divergence.

[wuming]

Hi, Charles

Did your case diverge finally?

[cgw]

Sorry, that was over 3 months ago and I’m not completely sure. We handle a lot of support requests and I don’t remember all the details of every one. Are you still having trouble with this simulation?

[wuming]

Yes, this problem hasn’t been solved yet. How long did your case run? According to my experience, it will diverge after several second.