[vExus]

Hi,
I wrote an UDF which fluent interprets without any errors but when initializing the solution i get:

Error: 
FLUENT received fatal signal (ACCESS_VIOLATION)
1. Note exact events leading to error.
2. Save case/data under new name.
3. Exit program and restart to continue.
4. Report error to your distributor.
Error Object: #f

I found that

C_T(cell, thread)

is the reason. I dont undestand why this happens because i used this macro while defining 3 different properties and only one time it couse this kind of problem.
Energy equation is turned on.
Whole code:

#include "udf.h"	

DEFINE_PROPERTY(gestosc_miesa, cell, thread)

{
	

	real rom;
	real t = C_T(cell, thread);
	
	
		rom = 1226 - 0,56 * t;

		return rom;
}




DEFINE_PROPERTY(wsp_przew_ciepla_mieso, cell, thread)

{
	

	real km;
	real t = C_T(cell, thread);
	
	
		km = 0,044 + t * 1,56 * 10^-3;

		return km;
}




DEFINE_PROPERTY(cp_miesa, cell, thread)

{
	real cpm;
	real m;
	real t = C_T(cell, thread); //here is the problem
	m = 0,7;
	
	if (t < 303.)
    		cpm = 1675 + 2510 * m;
  
  	else
    		cpm = 3000.;
		

		return cpm;
}










DEFINE_PROPERTY(gestosc_tluszczu, cell, thread)

{
	

	real rot;
	real t = C_T(cell, thread);
	
	
		rot = 1052 - 0,377 * t;

		return rot;
}

What should I do?

[blackmask]

It does not hurt to check the pointer before accessing it

real t;
if (NNULLP(THREAD_STORAGE(thread, SV_T))) 
  t = C_T(cell, thread);
else
  t = 300.0; /* or whatever appropriate */

[vExus]

It doesnt help...
I even tried this:

#include "udf.h"	
DEFINE_PROPERTY(cp_miesa, cell, thread)

{
	real t;

if (NNULLP(THREAD_STORAGE(thread, SV_T))) 
  t = C_T(cell, thread);

else
  t = 300.0; /* or whatever appropriate */
	

	return t;
}

just to check if it works and I got the same error.
What else can cause such an error?

Deleting "C_T(cell, thread)" from code makes error disappear.

[blackmask]

It does not make sense to me. Could you post your full code with the non-null-pointer checking?

[vExus]

Everything but

DEFINE_PROPERTY(cp_miesa, cell, thread)

{
	real cpm;
	real m;
	real t;
	m = 0,7;
	
	if (NNULLP(THREAD_STORAGE(thread, SV_T))) 
	{ 
		t = C_T(cell, thread);
		if (t < 303.)
    		cpm = 1675 + 2510 * m;
	}
	else
	{
			cpm = 3000.;
	}

	return cpm;
}

works fine.


Whole code:

#include "udf.h"	

DEFINE_PROPERTY(gestosc_miesa, cell, thread)

{
	

	real rom;
	real t = C_T(cell, thread);
	
	
		rom = 1226 - 0,56 * t;

		return rom;
}



DEFINE_PROPERTY(wsp_przew_ciepla_mieso, cell, thread)

{
	

	real km;
	real t = C_T(cell, thread);
	
	
		km = 0,044 + t * 1,56 * 10^-3;

		return km;
}



DEFINE_PROPERTY(cp_miesa, cell, thread)

{
	real cpm;
	real m;
	real t;
	m = 0,7;
	
	if (NNULLP(THREAD_STORAGE(thread, SV_T))) 
	{ 
		t = C_T(cell, thread);
		if (t < 303.)
    		cpm = 1675 + 2510 * m;
	}
	else
	{
			cpm = 3000.;
	}

	return cpm;
}




DEFINE_PROPERTY(gestosc_tluszczu, cell, thread)

{
	

	real rot;
	real t = C_T(cell, thread);
	
	
		rot = 1052 - 0,377 * t;

		return rot;
}




DEFINE_PROFILE (T_ot, thread, index)
{
	int s;
	real tval;
	double tab[8547]={}; 
	face_t f;
	
	s=RP_Get_Real("flow-time");
	tval = s < 8547 ? tab[s] : tab[8546];
	
	begin_f_loop(f, thread)
	{
		F_PROFILE(f, thread, index) = tval;
	}
	end_f_loop(f, thread);
}

[blackmask]

You have several "C_T" in your code and only this one is wrapped by pointer-checking. When you said that the error disappear after deleting the "C_T" line, did you delete this specific "C_T" line or did you delete all the occurrence of "C_T"? If this specific line caused segmentation fault, I have no idea what is going on but I doubt it should happen.

You are French, right? Note that 0,56 1,56 0,377 should be changed to 0.56, 1.56 and 0.377, respectively. The expression "1,56 * 10^-3" should be changed to "1.56e-3".

"

[vExus]

I know it seems strange but only this specific C_T couse trouble.

I am from Poland. Thanks for your advice.