Mass flow rate and conductive heat transfer in csp receiver

MFiX How to
#1

Hey,

so I’m trying to simulate mass flow rate and conductive heat transfer from obstructions to the particles in a DEM simulation.
I’ve set the wall temperature to be a 1000 K and I want to study the temperature gradient when the particle comes in contact with the obstruction.
it doesn’t show me any errors but it also wont run so I’m not sure what to fix.
I have attached the bug report below.

invertedvobstruction_2023-06-26T180446.637216.zip (801.6 KB)

#2

Running your job on Linux I see this:

shot-2023-06-26_19-10-05
shot-2023-06-26_19-10-05806×434 91 KB

I see from the bug report that you are on Windows … there was no error popup for you? If that’s the case, then we have two bugs to investigate

#3

No, when I try to run the program
it shows starting and then immediately resets.
so I cant really tell where the issue is.

#4

Running on Windows (with the 23.1.1 release) I get this popup:

Screenshot 2023-06-27 114727
Screenshot 2023-06-27 1147271012×512 39.3 KB

It doesn’t have as much detail as the one on Linux (full-stack) but this is expected - it’s due to limitations of the OS. But I would like to know why you are not getting the error popup at all.

#5

does this mean something is wrong with my stl file?

#6

Yes, that seems likely

#7

that seems to be the only error I’m getting. how do I fix that?

#8

@jeff.dietiker We’re seeing a 100% reproducible crash in stl_preproc_des_mod.f

   186           DO KK=K1,K2
   187           DO JJ=J1,J2
   188           DO II=I1,I2
   189              IF(dg_is_ON_myPE_plus1layers(II,JJ,KK)) THEN
   190                 IJK = DG_FUNIJK(II,JJ,KK)
   191                 CALL ADD_FACET_FOR_DES(II,JJ,KK,IJK,NN)
   192              ENDIF
   193           ENDDO
   194           ENDDO
   195           ENDDO

The value of ijk is way out of bounds:

(gdb) print ii
1047
(gdb) print jj
2686
(gdb) print kk
1331
(gdb) print ijk
-1104760546

Any idea what’s going on here?

#9

Is the geometry at the correct scale (over 300 meters) with the particles size of 3.5 cm?
The issue here is the des grid spacing is set to 3 times the particle diameter, which yields a humongous mesh of 16 Billion cells, larger than we can hold in memory (largest integer is about 2 Billion so we can’t even store the cell index).

DESGRIDSEARCH_IMAX =     3571
DESGRIDSEARCH_JMAX =     3380
DESGRIDSEARCH_KMAX =     1330

You can overwrite these values in the Solids>DEM pane (search grid partition) to smaller values, but the bottom line is you won’t be able to run this DEM simulation as is in a reasonable time. It has 81 Million particles.