[Mansoor_shad]

Hi,
I hope it is the right forum to ask. I want to simulate single DMFC using fluent 16.0 for study concentration in the flow channels of the bipolar plates and performance of the DMFC.
I'm writing a UDF to calculate current density in the fuel cell according to the following equation:

Icell=(1/Lcell)*(double integral over "ACL", ja*dxdy)

"I hope the formula is clear".

My c-code is:
DEFINE_ADJUST(i_cell,d)
{
Thread *t;
real i;
real x[ND_ND];
real sum_i=0.;
cell_t c;
thread_loop_c(t,d)
{
begin_c_loop(c,t)
C_CENTROID(x,c,t);
sum_i += C_UDMI(c,t,0)*C_VOLUME(c,t)/L_cell; /// C_UDMI(c,t,0)=ja/////
end_c_loop(c,t)
}
printf("Volume integral of current density: %g\n", sum_i);
}
1. Is my UDF code correct?
2. Should I use “DEFINE_ADUJST” to calculate the integral?
Thank you.

[pakk]

1. Try it. If it does what you want, it is correct.
2. Probably not. I think you want DEFINE_ON_DEMAND.

[Mansoor_shad]

Hello my friend,
Thank you for spending time thinking and answering my Question.
Be successful...

[serene6390]

Hi Mansoor,

Did you get your code work good?
I'm also trying to calculate integral along a face using data stored in UDM, but my code does not work for now. I checked my data in UDM, but I got 0 for I1_integ. I still couldn't find where cause the error.

Any reply will be appreciated.
Thanks.
Jing

Domain *d;

face_t f;


Thread *t;

Thread *t0;

cell_t c;

cell_t c0;


d = Get_Domain(1);

t = Lookup_Thread(d,7);


/* Loop over all faces on wall */



begin_c_loop(c,t)

{

c0 = F_C0(f, t);

t0 = THREAD_T0(t);

F_CENTROID(x,f,t);

y=x[1];

I1_integ+=C_UDMI(c,t0,0)/(C_UDMI(c,t0,0)+sigma)*dx;
I2_integ+=C_UDMI(c,t0,0)/(C_UDMI(c,t0,0)+sigma)*(zt1-y)*dx;

}

end_c_loop(c,t)

[`e`]

Originally Posted by serene6390

I'm also trying to calculate integral along a face using data stored in UDM, but my code does not work for now. I checked my data in UDM, but I got 0 for I1_integ. I still couldn't find where cause the error.

What is the variable dx? Where have you declared and initialised the variables I1_integ and I2_integ? Why not use the surface integrals already available within Fluent?