Johnson-Jackson boundary condition with cut-cell

MFiX
#1

Hi,

I am wondering whether with the new release it is possible to use the Johnson and Jackson boundary condition with the cartesian grid.
I’ve tried activating it in the sylane pirolysis tutorial, which is the only 2D tutorial with the cut-cell activated, and the simulation starts without any error message or warning, though it does not converge.
I am using the GUI version.

Thank you
Regards
Fabio

Spouted bed boundary region definition
Some questions about wall boundary condition about the solids phase in the tfm?
Moving wall for particle packing
#2

JJ boundary is currently only available for regular grid .

#3

Hi,

then why the existence of the subroutine “CG_CALC_GRBDRY” from line 1650 in calc_grbdry.f?
And why this piece of code in “CG_source_u_s.f”:
CASE(CG_PSW)
IF(BC_JJ_PS(BCV)==1) THEN ! Johnson-Jackson partial slip bc
NOC_US = .FALSE.
UW_s = BC_UW_S(BCV,M)
CALL CG_CALC_GRBDRY(IJK, ‘U_MOMENTUM’, M, BCV, F_2)
A_M(IJK,0,M) = A_M(IJK,0,M) - Area_U_CUT(IJK)*F_2
B_M(IJK,M) = B_M(IJK,M) - Area_U_CUT(IJK)F_2UW_s

Thanks
Fabio

#4

NOC_US = .FALSE. —> The effect of cut cells is not included!
UW_s = BC_UW_S(BCV,M)----> U velosity at wall is set to BC_UW_S(BCV,M), that is the value provided in the .dat or.mfix file.

Also, in SUBROUTINE CG_CALC_GRBDRY(IJK, TYPE_OF_CELL, M, L, F_2), see lines
1804 !THIS IS NOT CORRECT FOR EVERY DIRECTION**
1805 ! Calculate velocity components at i+1/2, j+1/2, k (relative to IJK1) ! not converted to CG

In file CG_source_u_s.f, see SUBROUTINE CG_SOURCE_U_S(A_M, B_M, M), the
line 425 : ! partial slip (WARNING:currently same as FSW)

Hope this will be helpful.

#5

Hi GaoXi,

So do you have any suggestions on the set of JJ boundary condition on the wall of a cylinder fluidized bed?
Thank you very much.

#6

Hi Fabio,

Have you solved this problems now? We also want to set JJ boundary condition on the wall of a 3d cylinder fluidized bed. Thank you very much.

#7

Option 1: You can use cylindrical coordinate. See the tutorial \legacy_tutorials\SpoutedBedCombustor
Personally, I do not recommend cylindrical coordinate as it was criticized in many papers for fluidized bed simulation.

Option 2: You can use regular cell to construct the cylindrical bed, like the old fashioned simulation many years ago when cut-cell are not available in MFiX.

Option 3: You can try to implement the JJ boundary for cut-cell.

Option 4: Coarse grid TFM (filtered TFM) simulation, which suggests free slip boundary for solid., thus can run with cut cell and without using JJ boundary.

#8

Hi GaoXi,

Thanks for your suggestions.
Hopefully, option 3 should be the best choice. Hope MFIX can implement such function in a near future.

#9

Hi Gao Xi,
Would you please explain the Option 2 more? How can we implement it?
I want to model a conical frustum shape FB. Is there anyway to model it without cut-cell method so that I can implement the JJ BC?

#10

Please see the tutoral case: legacy_tutorials\SpoutedBedCombustor

image
image279×777 27.5 KB

#12

I didn’t want to open new topic about this subject. Can’t we still use it for cut-cell?

#13

I think we cannot yet. However, it would be useful to use this BC with the cut-cell method.