[vincentztl]

Dear friends,

I'm trying to export flow rate as well as the adjacent cells ID for each face in the domain.
The first question:
My code can be compiled and loaded into FLUENT, but when I use "execute on demand" :
-----------------------------------------------------------------------------------
999999: mpt_accept: error: accept failed: No such file or directory

999999: mpt_accept: error: accept failedMPI Application rank 0 exited before MPI_Finalize() with status 2
The fl process could not be started.
-----------------------------------------------------------------------------------
The second question:
In UDF manual: Get_Domain(domain_id)
Does "domain_id" have the same meaning as "zone XXX" ?
e.g. 16000 hexahedral cells, zone 7, binary. zone 7 means domain_id is 7?




I'm a beginner and this is the first time I write UDF by myself. Maybe I made some stupid mistakes. Really hope someone can help me.


Here's my code:

#include "udf.h"
#include "mem.h"

 DEFINE_ON_DEMAND(getflux)
 {
 	FILE *fp;
 	Thread *t;
	Domain *domain= Get_Domain(1);
	face_t face;
 	cell_t c,c0,c1;
 	real fl;
 	int n;
 	
 	fp = fopen("data.txt","a+");
 	fprintf(fp,"c0,c1,flux\n");
	thread_loop_c(t,domain)
	{
		begin_c_loop(c,t)
		{
			c_face_loop(c,t,n)
			{
				face = C_FACE(c, t, n);
				c0 = F_C0(face,t);
				c1 = F_C1(face,t);
				fl = F_FLUX(face,t);
				
				fprintf(fp,"%d,%d,%g\n",c0,c1,fl);
			}
		}
		end_c_loop(c,t)
	 } 
 	fclose(fp);
 }

Thank you!
Ps.My English is not good, I'm sorry if it's bothering you.

[arunraj]

Try this.

#include "udf.h"
#include "mem.h"

DEFINE_ON_DEMAND(getflux)
{
Thread *t;
Domain *d;
d= Get_Domain(1);
face_t f;
cell_t c,c0,c1;
real fl;
int n;

t = Lookup_Thread(d, PROVIDE_THE_ID_HERE);
thread_loop_c(t,d)
{
begin_c_loop(c,t)
{
c_face_loop(c,t,n)
{
face = C_FACE(c, t, n);
c0 = F_C0(f,t);
c1 = F_C1(f,t);
fl = F_FLUX(f,t);
}
}
end_c_loop(c,t)
}
}

[AlexanderZ]

your code seems to be correct for single core computation,
the only thing could be:
was

 int n;

to be

int n=0;

if you want to write into file in parallel fluent, you need to modify your code
Check ANsys Fluent Customization manual for more details

[vincentztl]

Hi arunraj and AlexanderZ,
Thanks for your advice !
I've tried your code but the program still can't run successfully. Thank you anyway.
Fortunately, I figure it out. Seems like it's the problem with the using of pointers. The following code can run successfully. Hope it help someone someday.

#include "udf.h"
#include "mem.h"

 DEFINE_ON_DEMAND(getflux)
 {
 	FILE *fp;
 	Thread *c_thread,*f_thread;
	Domain *domain= Get_Domain(1);
	face_t face;
 	cell_t c,c0,c1;
 	real fl;
 	int n;
 	
 	fp = fopen("data.txt","a+");
 	fprintf(fp,"c0,c1,flux\n");
	thread_loop_c(c_thread,domain)
	{
		begin_c_loop(c,c_thread)
		{
			c_face_loop(c,c_thread,n)
			{
				face = C_FACE(c, c_thread, n);
				f_thread=C_FACE_THREAD(c,c_thread,n);
				c0 = F_C0(face,f_thread);
				c1 = F_C1(face,f_thread);
				fl = F_FLUX(face,f_thread);
				fprintf(fp,"%d,%d,%g\n",c0,c1,fl);
			}
		}
		end_c_loop(c,c_thread)
	 } 
 	fclose(fp);
 }