[scien]

Hi,
When I calculate the derivatives of a UDS, the result is very strange at boundaries. I used C_UDSI_G.
The values are correct inside the domain though.

I have tested this with a very simple case.
If I set the value of UDS as a constant, the derivative should be zero but at boundary it has some values.
Even though they are very small values, they can affect other calculations when they become denominator.

Why does this happen? Is there any solution for this?

[jc2006]

I have found the similar problem too. Maybe you can calculate the gradient by your code

[Bruno Machado]

This is a bug of Fluent. I contacted their support and it was reported to the programmers team. What you can do to decrease the effect of it is to set that if the cell is in the boundary, the gradient is equal to 0.

The following code is the one I used in my problem to set it as 0, might help you guys.

DEFINE_EXECUTE_AT_END(gradient_zero_boundary)
{
Domain *d;
face_t f;
cell_t c, c0, c1;
Thread *t, *t0, *ct, *t1; 
d = Get_Domain(1);
	
anode_diffusion_shadow_id = RP_Get_Integer("fuel-cell/anode-diffusion-shadow-id");
cathode_diffusion_shadow_id = RP_Get_Integer("fuel-cell/cathode-diffusion-shadow-id");



/* Initialise Cells */
thread_loop_c(ct, d)
{
	begin_c_loop(c,ct)
	{
		C_UDMI(c,ct,elec_x) = C_UDSI_G(c,ct,ELEC_POT)[0];
		C_UDMI(c,ct,elec_y) = C_UDSI_G(c,ct,ELEC_POT)[1];
		C_UDMI(c,ct,elec_z) = C_UDSI_G(c,ct,ELEC_POT)[2];
	}
	end_c_loop(c,ct)
}

/* Loop over all faces on wall */

t = Lookup_Thread(d,anode_diffusion_shadow_id);
	
	begin_f_loop(f,t)
	{
	c0 = F_C0(f, t);
	t0 = THREAD_T0(t);
	
		C_UDMI(c0,t0,elec_x) = 0.0;
		C_UDMI(c0,t0,elec_y) = 0.0;
		C_UDMI(c0,t0,elec_z) = 0.0;
	}
	end_f_loop(f,t)

	
t = Lookup_Thread(d,cathode_diffusion_shadow_id);

	begin_f_loop(f,t)
	{
	c1 = F_C1(f,t);
	t1 = THREAD_T1(t);
	
		C_UDMI(c1,t1,elec_x) = 0.0;
		C_UDMI(c1,t1,elec_y) = 0.0;
		C_UDMI(c1,t1,elec_z) = 0.0;
}
end_f_loop(f,t)
}

[scien]

Thanks, David.
I would like to do it but I am not so good at programming.
Could you let me know it if you don't mind?

Also Thanks for your code, Machado. But I have to use the gradient on boundaries and I can't set it as 0.
Thanks anyway.

[jc2006]

Hi Machado
I think, maybe one can calculate the gradient of UDS by the information of the two adjacent cell on boundary. I find that you are modeling fuel cell, and I am doing the similar work too. Hope we can have more communication in the future

[jc2006]

Hi Hugo
In my case, the gradient of UDS is directly calculated by (C_UDSI(c0,t0,i)-C_UDSI(c1,t1,i))/(X(1)-X(0))(take x-direction for example).

[scien]

Hi David.
I will apply it though I am a little worried about how I can modify it on the boundary.
Thanks again.