get "nan" on energy is iterated from converged cold-flow
The basic procedure for underhood case is:
1. activate SKE, DO, and Energy models
2. setup boundary conditions for flow, temperature, and radiation in serial solver (includes turning on planar conduction)
3. partition case while in serial and save case
4. read in parallel
5. select and converge only cold flow (1st-order and then 2nd-order)
6. de-select flow equations and select DO and energy
7. converge DO and energy (1st-order and then 2nd-order)
This procedure was working fine in v5.x. But in v6.0, before 1st iteration in Step 5, it would give the following error message:
Error: > (greater than): invalid argument : wrong type [not a number]
Error Object: nan
While this is under investigation, what should be done for workaround?
1. read case and data in serial
2. Turning Off Planar Conduction
o- write out bc file by typing the following in TUI:
o- Edit bc file (e.g. bc-file) and replace
(planar-conduction? . #t)
(planar-conduction? . #f)
You can use VI or any text editor (e.g. nedit, dtpad) that has 'replace' functionality.
o- save the bc file under new name (bc-file-2)
o- read back the bc file:
/file read-bc bc-file-2
It will delete all the planar conduction zones.
Turn off radiation and energy
4. save case and data under new name
5. exit fluent
6. read back the case nad data files in serial
7. Turn on Planar Conduction
o- read the original bc file
8. Turn on radiation and energy
Note that the thickness, emissivities, and temperature boundary conditions are already saved in the case.
9. Partition case
10. Save case and data
12. Read in parallel and iterate.
It should run.
NOTE: Some thermal data seems to get "corrupted". But by turning OFF and ON energy and DO, it "refreshes" it - it reinitializes the intensities and temperature field.
But why go through all the trouble? We could as well use the patch capabilities to patch temperature! The answer is that the patch is not applied at planar conduction zones.