FDS Alignement Problems for Meshes of Different Sizes

[gabferrara]

Hello.
I''m trying to exploit the Pyrosim capability of creating abutting meshes of different size.
I created 5 meshes with the schematic of the attached pdf file.
Each mesh perfectly adheres each other and the cells of the finer mesh are integer multiple of the cells of the coarser meshes.
Pyrosim is telling me that everything is fine with the mesh alignment test (all are "passed").
But when I run FDS, it tells me that Mesh-1 cannot align with Mesh-3 etc... and the simulation cannot start.
Would you please give me some help.
 

[Charlie Thornton]

We''ve tried to reproduce the same kind of alignment tests in PyroSim that are run by FDS, but it''s possible that we''ve either missed a case or there''s a difference in tolerance.

The question is if your mesh #3 has the alignment problem shown in the attached image (from figure 6.2 of the FDS manual) when compared with any of the other meshes.  Could you post your &MESH FDS input records so I could see the numbers?

I really want to add an option that lets you enter a bounding box and do an automated mesh decomposition and refinement (split a single mesh into 9).  I think that would make cases like this a lot more straightforward.

[gabferrara]

First of all many thanks for yor ready reply.
Please find the psm file attached.

Gabriele

[Charlie Thornton]

There appears to be a limit to how big of a cell size change FDS will tolerate.  If the central column is finer by a factor of up to 19 it runs, at 20 it fails the alignment test. 

Here is the FDS code that checks:

Code:

      MM=>MESHES(NOM)
      ALIGNED = .TRUE.
      IF ( (ABS(IOR)==2 .OR. ABS(IOR)==3) .AND. MM%DX(IIO_MIN)<=M%DX(I) .AND. &
            ABS( ((IIO_MAX-IIO_MIN+1)*MM%DX(IIO_MIN)-M%DX(I)) / MM%DX(IIO_MIN))>0.05 ) ALIGNED = .FALSE.
      IF ( (ABS(IOR)==1 .OR. ABS(IOR)==3) .AND. MM%DY(JJO_MIN)<=M%DY(J) .AND. &
            ABS( ((JJO_MAX-JJO_MIN+1)*MM%DY(JJO_MIN)-M%DY(J)) / MM%DY(JJO_MIN))>0.05 ) ALIGNED = .FALSE.
      IF ( (ABS(IOR)==1 .OR. ABS(IOR)==2) .AND. MM%DZ(KKO_MIN)<=M%DZ(K) .AND. &
            ABS( ((KKO_MAX-KKO_MIN+1)*MM%DZ(KKO_MIN)-M%DZ(K)) / MM%DZ(KKO_MIN))>0.05 ) ALIGNED = .FALSE.
      IF (.NOT.ALIGNED) THEN
         WRITE(LU_ERR,''(A,I3,A,I3)'') ''ERROR: MESH '',NM,'' is out of alignment with MESH '',NOM
         PROCESS_STOP_STATUS = SETUP_STOP
         IERR = 1
         RETURN
      ENDIF

I didn''t bother to track down all those variables, but that tolerance of 0.05 matches pretty well with the 1/20th I mentioned above.  It looks like you can''t have a cell size difference of more than a factor of 19 in any direction.

[gabferrara]

Yes you''re right: it works with a smaller cell size ratio.
This is very powerful for my applications to pool fires in the open.
I need fine grids where the fire is burning and coarser where the gradients have smeared out.
Thank you again.

Gabriele

[Bryan Klein]

Gabriele,

I wanted to add to the discussion that mesh difference ratios of greater than 4:1 are typically not recommended.  It may be better to nest the meshes within other meshes that are less coarse, building up to the most coarse mesh.  20:1 or more is far away from the recommended values.

See this FDS-SMV Discussion Group thread for a reference to this recommendation by Randy McDermott.
http://groups.google.com/group/fds-smv/msg/cc62c615573126c7

Also, without the new embedded mesh feature of FDS_6 the coarse mesh only provides background information to the more refined mesh, but the refined mesh does not pass information back to the coarse mesh.  The fine mesh and the coarse mesh are two different calculations.  In the new version, the refined mesh will be the dominating solution in that volume and will pass that data into the coarse mesh cells.

Best Regards,
-Bryan Klein

[gabferrara]

Hello Bryan.

Thank you for your reply.
I am very interested in this topic.
Can I ask you a specific issue of aligned meshes.
Say I have a refined grid around a pool fire and less refined grids embedding the finer mesh.
How reliable is the calculated intensity of heat radiation passed from the fine grid to the coarse grid around it.
I thank you in advance for your support.
regards.

Gabriele

Gabriele,

I wanted to add to the discussion that mesh difference ratios of greater than 4:1 are typically not recommended.  It may be better to nest the meshes within other meshes that are less coarse, building up to the most coarse mesh.  20:1 or more is far away from the recommended values.

See this FDS-SMV Discussion Group thread for a reference to this recommendation by Randy McDermott.
http://groups.google.com/group/fds-smv/msg/cc62c615573126c7

Also, without the new embedded mesh feature of FDS_6 the coarse mesh only provides background information to the more refined mesh, but the refined mesh does not pass information back to the coarse mesh.  The fine mesh and the coarse mesh are two different calculations.  In the new version, the refined mesh will be the dominating solution in that volume and will pass that data into the coarse mesh cells.

Best Regards,
-Bryan Klein