Learning Forum

[dseong]

Hi all,

I am working on creating a UDF to track particles hitting the outlet boundary. The setup involved in an air purifier in a small chamber, and I aim to set the purifier's efficiency at 95%. (95% of the particles that reach the outlet will escape, while 5% will be reinjected from a reinjection point)

From my understanding:
return PATH_END is used for considering particles escaped from the domain and
return PATH_ACTIVE is for continuously tracking particles

I would like to set up the system so that 5% of particles are reinjected from the reinjection point and continue being tracked. While PATH_END seems to work as expected, PATH_ACTIVE does not. In the console, particles are being aborted.

I have revised my code several times and attempted different solutions, but I am currently out of ideas.

Would anyone be able to help me solve this issue or suggest any possible approaches? I would greatly appreciate your insights.

Thank you in advance!

Best regards,
Dahae

--------------------------------------------------------------
Here is my code FYI:

#include "udf.h"
#include "dpm.h"

#define REINJECT_POSITION_X 2.0 // Set the x-coordinate of the reinjection point
#define REINJECT_POSITION_Y 0.425 // Set the y-coordinate of the reinjection point
#define REINJECT_POSITION_Z 0.18 // Set the z-coordinate of the reinjection point

/* Global variables */
int particle_count = 0;
int escaped_count = 0;
int reinjected_count = 0;

DEFINE_DPM_BC(udf_outlet, tp, t, f, f_normal, dim)
{
particle_count++;

if (particle_count <= 95)
{
/* Mark as escaped */
escaped_count++;

return PATH_END;

}

if (particle_count > 95)
{
/* Mark as reinjected */
reinjected_count++;

/* Move particle to specific point */
P_POS(tp)[0] = REINJECT_POSITION_X; // Specific X position
P_POS(tp)[1] = REINJECT_POSITION_Y; // Specific Y position
P_POS(tp)[2] = REINJECT_POSITION_Z; // Specific Z position

return PATH_ACTIVE;
}

/* Reset the counter after 100 particles */
if (particle_count > 100)
{
particle_count = 0;
escaped_count = 0;
reinjected_count = 0;
}

return PATH_ACTIVE;
}

[Rob]

Are you applying the UDF to the outlet boundary? What fate did you set on the boundary? 

[dseong]

Hi Rob, thanks for your response.

Yes. I am applying the UDF to my outlet boundary (pressure outlet). This is what I set on the boundary.

[Rob]

How far into the domain is the injection?

[dseong]

 

I injected around 10,000 particles uniformly all around the domain (the size of the domain is 3 m by 4 m by 2.5 m).  And I wanted to set the 5% of particles passing the pressure outlet to be reinjected from the inlet. So, I specified the reinjection point in front of the inlet. 

 

[Rob]

But you're not sending 5% back you're sending count-95 back. Or would be if it worked. 

[dseong]

From my understanding, 'return PATH_END' means that particles escaped from the domain. So, I thought my code said that the particle count of less than 95 escaped. Could you please correct my code? 

if (particle_count <= 95)
{
/* Mark as escaped */
escaped_count++;

return PATH_END;

}

if (particle_count > 95)
{
/* Mark as reinjected */
reinjected_count++;

/* Move particle to specific point */
P_POS(tp)[0] = REINJECT_POSITION_X; // Specific X position
P_POS(tp)[1] = REINJECT_POSITION_Y; // Specific Y position
P_POS(tp)[2] = REINJECT_POSITION_Z; // Specific Z position

return PATH_ACTIVE;
}

[Rob]

I'm not debugging, the problem is if (particle_count > 95) I think. 

[dseong]

Thanks for your reply, Rob

I am sorry but I am having trouble understanding the issue. If the problem lies with particle_count>95, how should I relocate particles from 96th to 100th that are hitting the outlet? I would greatly appreciate your answer. 

[Rob]

If you have 100 parcels you'll have the last 5 be reinjected. But... with 1000 particles you'll have 905 reinjected. So a random number function may be better. 

Now, for the actual return part. If a parcel hits the outlet does it leave the domain before the UDF triggers? If you put a surface at a few cells into the domain (porous jump are really useful for this) does behaviour change?