[C_Zhang]

Hi everyone. I am writing a udf for three-phase-flow.After I hooked it to my model,it initialised well but can't calculate.Below are my udf.
#include"udf.h"
#include"unsteady.h"

DEFINE_ADJUST(store_gradient,d)
{
real unit_normal;
Thread *big_thread,*mixture_thread,*small_thread;
cell_t cell;
int current_iter,domainid,threadid1,threadid2;
current_iter=(nres==0)?(0)(int)count2[nres-1]);
domainid=8;
threadid1=1;
threadid2=2;
mixture_thread=Lookup_Thread(d,domainid);
big_thread=THREAD_SUB_THREAD(mixture_thread,thread id1);
small_thread=THREAD_SUB_THREAD(mixture_thread,thre adid2);
if((current_iter==0)&&(CURRENT_TIME==0))
{
begin_c_loop(cell,big_thread)
{
C_UDMI(cell,big_thread,0)=0;
C_UDMI(cell,big_thread,1)=0;
C_UDMI(cell,big_thread,2)=0;
C_UDMI(cell,big_thread,3)=0;
C_UDMI(cell,big_thread,4)=0;
}
end_c_loop(cell,big_thread)
begin_c_loop(cell,small_thread)
{
C_UDMI(cell,small_thread,5)=0;
C_UDMI(cell,small_thread,6)=0;
C_UDMI(cell,small_thread,7)=0;
C_UDMI(cell,small_thread,8)=0;
C_UDMI(cell,small_thread,9)=0;
}
end_c_loop(cell,small_thread)
}
if (!Data_Valid_P())
return;
begin_c_loop(cell,big_thread)
{
C_UDSI(cell,big_thread,0)=0;
C_UDSI(cell,big_thread,1)=0;
C_UDSI(cell,big_thread,2)=0;
}
end_c_loop(cell,big_thread)
Message("helloworld");
return;
}

real integrand(real scale,real density,real diameter,real tension,real dissipation)
{
real answer;
answer=exp(-12*diameter*diameter*tension/(density*2.047*pow(scale,11.0/3.0)*pow(dissipation,2.0/3.0)))/pow(scale,4.0);
return answer;
}
real integrand1(real scale,real density,real diameter,real tension,real dissipation)
{
real answer;
answer=exp(-12*diameter*diameter*tension/(density*2.047*pow(scale,11.0/3.0)*pow(dissipation,2.0/3.0)))/pow(scale,11.0/3.0);
return answer;
}
DEFINE_PROPERTY(diameter,c,t)
{
real diameter1,surface_tension,density_of_continuous,vi scosity,voidreturn;
Thread *mixture_thread,*big_thread;
int threadid;
threadid=1;
density_of_continuous=998.2;
surface_tension=0.0719404;
viscosity=0.001003;
mixture_thread=THREAD_SUPER_THREAD(t);
big_thread=THREAD_SUB_THREAD(mixture_thread,thread id);
voidreturn=1;
if(C_VOF(c,t)==0)
diameter1=0.0008;
else
{
diameter1=1.78*pow(sqrt(surface_tension/((density_of_continuous-(C_R(c,big_thread)*C_VOF(c,big_thread)+C_R(c,t)*C_ VOF(c,t))/(C_VOF(c,big_thread)+C_VOF(c,t)))*9.81)),0.889)*po w(viscosity/density_of_continuous,0.0836)/pow(C_D(c,mixture_thread)*C_VOF(c,t)/(C_VOF(c,big_thread)+C_VOF(c,t)),0.0279);
}
return (diameter1);
}
Attachment is what I did in user defined scalar dialog box. All my three scalars have the same setting.

[C_Zhang]

My udf seems strange because it's too long and all parts excluded in this webpage have been set between /* and */ so they won't execute.

[C_Zhang]

I am sure the problem is on UDS because when I set 42nd,43rd and 44th line between /* and */, it works well.
Those lines are
C_UDSI(cell,big_thread,0)=0;
C_UDSI(cell,big_thread,1)=0;
C_UDSI(cell,big_thread,2)=0;

[C_Zhang]

I will appreciate any of your help

[AlexanderZ]

Hello,

you are using multi-phase flow, do I not know how to start such cases properly.
However, you are using begin_c_loop(cell,big_thread) and begin_c_loop(cell,small_thread) to set/calculate variables for specific regions.
But what values do these variable have at other regions?

I did not test it, but what if UDS values should be defined everywhere?
In that case you need loop over all threads to initialize UDS values, but then you may modify values of UDS variables in specified regions.

I hope my idea is clear.

Best regards,
Zorin Alexander

[C_Zhang]

Dear Zorin
Thank you for your help first.
UDS is only used to calculate the second order gradient of void fraction of each secondary phase in my UDF. Do you have any suggestions?

[AlexanderZ]

Try to use loop over all threads for UDS for initialization

cell_t cell;
Thread *thread;
thread_loop_c (thread,domain)
{
		begin_c_loop (cell,thread)
               {
                   C_UDSI(cell,thread,0)=0.0;
		end_c_loop (cell,thread)
	       }
}

[AlexanderZ]

or you may try to use big_thread directly, however, I have never tried it before:

thread_loop_c (big_thread,domain)
{
		begin_c_loop (cell,big_thread)
               {
                   C_UDSI(cell,big_thread,0)=0.0;
		end_c_loop (cell,big_thread)
	       }
}

[C_Zhang]

I'm sorry maybe I gave you too little information.
Actually, I want to get the second order gradient of void fraction for each secondary phase, but I don't know how to do. This way is found from some predecessors' threads here in this forum. Since I don't know its detail, I may have made some mistakes on using UDS as well as on setting something in GUI. Do you know some effective ways to get second order gradient?
Thank you very much

[C_Zhang]

Oh! your reply is so quick, I should read it first

[C_Zhang]

looping over all threads in subdomain works! But I am still confused. Why do I have to loop over all threads in subdomain?

[AlexanderZ]

Originally Posted by C_Zhang

looping over all threads in subdomain works! But I am still confused. Why do I have to loop over all threads in subdomain?

Frankly speaking, I do not know,
may be the reason is that fluent tries to calculate scalar everywhere, so without initialization it gives error

Best regards

[C_Zhang]

Thank you.Fluent is a mysterious software