[Eanna.Kennedy]

I am trying to calculate the standard deviation of particles about the centre of a channel at a sampling plane near the exit and I am having trouble getting my UDF to work. The UDF compiles and loads fine, but it causes fluent to crash when i try to run it. I have included my code below. I think the problem is probably due to the need to loop through all particles in a Macro that doesn't take in any DPM arguments. Any advice would be really appreciated. Thanks.

#include "udf.h"
#include "dpm.h"
#include "surf.h"
#include "dpm_types.h"
#include "dpm_laws.h"

/************************************************** ****************/
/* UDF records Standard Deviation of all particles on a plane */
/************************************************** ****************/

double Standard_Deviation = 0.0; /* intitialise standard deviation as global variable*/

/*Creates Output Parameter Standard Deviation*/
DEFINE_REPORT_DEFINITION_FN(standard_deviation)
{
return Standard_Deviation;
}

DEFINE_ON_DEMAND(Get_Particle_Positions)
{
double d = 0.0;
double y = 0.0;
double z = 0.0;
double dbar = 0.0;
double deviation = 0.0;
double counter = 0.0;
int zone_ID = 3; /* initialise zone id for sampling plane*/
Thread *c_thread; /* c_thread is declared as a variable */
Domain *domain; /* domain is declared as a variable */
Tracked_Particle *tp; /* tp declared as identifier for tracked particles */
cell_t c; /* c declared as identifier for cells */

domain = Get_Domain(1); /* returns fluid domain pointer */
c_thread = Lookup_Thread(domain,zone_ID);
tp = 0;

/* Loop calculates sum of particle-to-centre distances, allows mean distance to be calculated later*/

begin_c_loop(c, c_thread) /* loops over cells in a cell thread */
{
while(NULLP(tp)){ /* checks that tp points to actual particle*/
y = TP_POS0(tp)[1];
z = TP_POS0(tp)[2];
d += sqrt((y*y) + (z*z));
tp ++;
counter ++;
};
}
end_c_loop(c, c_thread)

dbar = d / counter; /* calculate mean particle-to-centre distance*/
tp = 0; /* reset particle pointer*/

/* Loop to calculate sum of squared particle-to-mean distances */
begin_c_loop(c, c_thread)
{
while(NULLP(tp)){ /* checks that tp points to actual particle*/
y = TP_POS0(tp)[1];
z = TP_POS0(tp)[2];
d = sqrt((y*y) + (z*z));
deviation += pow((dbar - d), 2);
tp ++;
};
}
end_c_loop(c, c_thread)

Standard_Deviation = deviation / counter;
}

[obscureed]

Hi Eanna,


I don't see in your code anything that finds a particle-pointer and puts it into *tp. So, you initialize "tp = 0;", and then test whether it is a NULLP pointer. If it *is* a null pointer, you go ahead and use it -- I would have thought the opposite would make more sense, although then your code would then (correctly) report that it found no particles.


You need to find a way to loop over DPM parcels. As it turns out, parcels' connection to fluid cells is one-way: parcels know which cell they are in, but cells do not know which parcels they contain. (This is not always true. For some spray models, and for DEM, you might find a loop called begin_particle_cell_loop. Typically, though, the information required for this loop is not maintained, or even allocated.)


So, please go and look for a particle loop, probably involving "loop(p,I->p)" and "Get_dpm_injections". Sorry, I don't have a good reference to recommend at the minute.



Good luck!
Ed

[Eanna.Kennedy]

Hi Ed,

Thanks for the reply!

I have changed the code significantly since I posted originally and it calculates the standard deviation of particles correctly. However I plan to use direct optimization to minimise the the standard deviation of particles near the exit plane.

I tried your suggestion of using loop(p, p->I) but it appears that it couldn't access the particle information unless I ran a dpm sample report at first.

At the moment I am using a DEFINE_DPM_OUTPUT macro to save particle positions to an array that I then use for calculations inside a DEFINE_REPORT_DEFINITION_FN. This means that I need to run a dpm sample report for every design point. Do you have any advice on how I could automate this? Or how i can access the necessary particle information without having to run a sample report first?

Thanks again,
Eanna

[djamila77]

Originally Posted by Eanna.Kennedy

I am trying to calculate the standard deviation of particles about the centre of a channel at a sampling plane near the exit and I am having trouble getting my UDF to work. The UDF compiles and loads fine, but it causes fluent to crash when i try to run it. I have included my code below. I think the problem is probably due to the need to loop through all particles in a Macro that doesn't take in any DPM arguments. Any advice would be really appreciated. Thanks.

#include "udf.h"
#include "dpm.h"
#include "surf.h"
#include "dpm_types.h"
#include "dpm_laws.h"

/************************************************** ****************/
/* UDF records Standard Deviation of all particles on a plane */
/************************************************** ****************/

double Standard_Deviation = 0.0; /* intitialise standard deviation as global variable*/

/*Creates Output Parameter Standard Deviation*/
DEFINE_REPORT_DEFINITION_FN(standard_deviation)
{
return Standard_Deviation;
}

DEFINE_ON_DEMAND(Get_Particle_Positions)
{
double d = 0.0;
double y = 0.0;
double z = 0.0;
double dbar = 0.0;
double deviation = 0.0;
double counter = 0.0;
int zone_ID = 3; /* initialise zone id for sampling plane*/
Thread *c_thread; /* c_thread is declared as a variable */
Domain *domain; /* domain is declared as a variable */
Tracked_Particle *tp; /* tp declared as identifier for tracked particles */
cell_t c; /* c declared as identifier for cells */

domain = Get_Domain(1); /* returns fluid domain pointer */
c_thread = Lookup_Thread(domain,zone_ID);
tp = 0;

/* Loop calculates sum of particle-to-centre distances, allows mean distance to be calculated later*/

begin_c_loop(c, c_thread) /* loops over cells in a cell thread */
{
while(NULLP(tp)){ /* checks that tp points to actual particle*/
y = TP_POS0(tp)[1];
z = TP_POS0(tp)[2];
d += sqrt((y*y) + (z*z));
tp ++;
counter ++;
};
}
end_c_loop(c, c_thread)

dbar = d / counter; /* calculate mean particle-to-centre distance*/
tp = 0; /* reset particle pointer*/

/* Loop to calculate sum of squared particle-to-mean distances */
begin_c_loop(c, c_thread)
{
while(NULLP(tp)){ /* checks that tp points to actual particle*/
y = TP_POS0(tp)[1];
z = TP_POS0(tp)[2];
d = sqrt((y*y) + (z*z));
deviation += pow((dbar - d), 2);
tp ++;
};
}
end_c_loop(c, c_thread)

Standard_Deviation = deviation / counter;
}

Thank you
can you help me to write the Nusselt UDF for the following formula
Nu=\frac{-1}{q_{cond}}\int_{0}^{L}\frac{\partial T}{\partial y}dx

[djamila77]

Hi there,
please I need a help

I need to know how I can applied the coupled and the uncoupled cas for the natural convection and radiation by fluent