[Aristid]

Hi everyone,

i tried to get the temperature variable via a User CEL Function by using USER_GETVAR.

However, when I call the function it doesn't return anything, even the CRESLT variable that should tell if the function call was successful is empty.

Then i simplified the routine and tried to get the pressure variable (mabye this function can't get the temperature variable?) but the same error accured.
The Code is down below.

How do i use this function properly? Is there a bug? If so, is there a workaround?





#include "cfx5ext.h"
dllexport(adt)
SUBROUTINE ADT (
& NLOC,NRET,NARG,RET,ARGS,
& CRESLT,
& CZ,DZ,IZ,LZ,RZ)

#include "MMS.h"
#include "stack_point.h"

INTEGER NLOC, NRET, NARG
CHARACTER*(*) CRESLT
REAL RET(NLOC,NRET), ARGS(NLOC,NARG)

INTEGER IZ(*)
CHARACTER CZ(*)*(1)
DOUBLE PRECISION DZ(*)
LOGICAL LZ(*)
REAL RZ(*)

__stack_point__ pT

CRESLT = 'GOOD'

CALL SET_A_0 (RET, NLOC*NRET )

CALL MESAGE ('WRITE', 'WORKS UNTIL SET_A0')

CALL USER_GETVAR ('Steelliq.pressure',CRESLT,
& pT,CZ,DZ,IZ,LZ,RZ)

IF (CRESLT .NE. 'GOOD') THEN
CALL MESAGE ('WRITE', 'GET_VAR NOT GOOD')
ENDIF

CALL MESAGE ('WRITE', CRESLT)

CALL USER_SOURCE_SUB (RET(1,1),ARGS(1,1))

END

SUBROUTINE USER_SOURCE_SUB (OUTP, INP)
REAL OUTP, INP
OUTP = 1000.0
CALL MESAGE ('WRITE', 'WORKS UNTIL USER_SOUCE')

END

[Opaque]

If I recall correctly, you can use

'Steelliq.Pressure'

or

'Steelliq.p'

but definitely NOT

'Steelliq.pressure'

What surprises me is that you are not getting CRESLT = 'NAME' as documented.

[Aristid]

Thank you very much for the quick reply!

I now corrected it to 'Steelliq.Pressure', and i also tried 'Steelliq.p',
but still the same error accurs (solver crashes and message "solvermanager.exe stopped working" appears).

Before that i also tried to get Variables via creating Additional variables
(someone suggested this in a master thesis online to avoid this "bug"),
but the call to USER_GETVAR fails every time and CRESLT is empty.


Have you used USER_GETVAR successfully?

Are there alternatives to using USER_GETVAR?

I've been stuck on this problem for over a week now and suspected this might be a bug.

[Opaque]

Have you run the tutorial or documentation example using user_getvar?

Would you mind posting the basics of your usage of user_getvar?

[Aristid]

Thank you very much! I have now run "User CEL Example 2" (AdVarSource), and it's running fine.

It looks like the error i experienced is caused by the time the Call to the User CEL Function is made.

Before i always tried to check the return Value of the User CEL Function via Monitor Objects but never actually used it.
Mabye if i use it somewhere the error will disappear.


My Goal is to find the Y-Coordinate of the position where Temperature has a certain Value (f.e. 2000K).

My Usage of user_getvar would be to obtain the temperature field in the entire domain (or mabye only on one surface, i don't know yet).

By usind user_get_meshdata i would obtain vertex coordinates.

Then i would find the index of the temperature value closest to 2000K and hope that the Y-Coordinate-Array has the same indexing as the Temperature Array.

[Opaque]

Great!!

You are aware your goal is not a 1->1 match, there can be multiple locations for a given value of Temperature. Not even sorting the arrays you are dealing with will guarantee you a unique solution, or even a repeatable solution, i.e. the same location between calls even if the temperature at the specific location remains constant.

Let us know how we can help