[agarhaji]

Hello everyone.

I am testing different drag correlations for non-spherical particles and have made this UDF and compiled it in Fluent.

The problem i am facing now is, that i want to plot the drag coefficient for the particles and drag force calculated from the UDF.

I have tried the user-defined memory but without any luck. The results i get are all zeros.

Can anyone help me with this one. Maybe i need something in my UDF?

#include "udf.h"
#include "dpm_types.h"


DEFINE_DPM_DRAG(drag_force,Re,tp)
{
double C_D, drag_force;
double d_sv, L, W, d_Cl, d_CC, Psi;
d_sv = TP_DIAM(tp); /* Sieve Diameter defined as the injected particle diameter */
L = 0.0115*log(d_sv) + 0.0818;
W = -42.365*pow(d_sv, 2) + 1.5702*d_sv - 0.0012;
d_Cl = W; /* Defining the diameter of the inscriped circle as the with of the particle*/
d_CC = L; /* Defining the diameter of the circumscribed circle as the with of the particle*/
Psi = sqrt(d_Cl / d_CC);
/* Calculating the drag coefficient using Yow's correlation */
C_D= (15.21 + (10.82 / Psi) - (0.14 / pow(Psi, 2))) / Re + (13.41 - (10.64 / Psi) - (0.016 / pow(Psi, 2))) / sqrt(Re) - 8.82 + 5.70 / Psi + 0.23 / pow(Psi, 2);
/* Calculating the dragforce as a dimentionless group */
drag_force = (18 * C_D*Re) / 24;
return (drag_force);
}

[bhargavbharathan]

Originally Posted by agarhaji

Hello everyone.

I am testing different drag correlations for non-spherical particles and have made this UDF and compiled it in Fluent.

The problem i am facing now is, that i want to plot the drag coefficient for the particles and drag force calculated from the UDF.

I have tried the user-defined memory but without any luck. The results i get are all zeros.

Can anyone help me with this one. Maybe i need something in my UDF?

Hi,

Firstly make sure you go to=>

Parameters & Customization -> User Defined Memory -> Number of User-Defined Memory Location and select the no. of UDMs you need.

Note that the UDM numbering starts from 0. So if you select 3 UDMs, their numbers would be 0, 1 and 2.

In your UDF you need to then assign your variable to a specific UDM. It would go something like:

F_UDMI(f,t,0)=drag_force;
C_UDMI(c0,t0,0)=drag_force;

Check out the below link:
https://www.sharcnet.ca/Software/Flu...udf/node96.htm


-B

[obscureed]

Hi Agarhaji,

As Bhargavbharathan points out, one option is to allocate User-Defined Memory (UDM) and save something in the cells containing the DPM parcels at any instant. This could be a single value for the latest parcel to pass through that cell, or you could (using more than one UDM) save average values. You could plot these by contour plots on section planes, etc. It would be a good idea to have a DEFINE_ON_DEMAND function to reset the UDM back to zero everywhere. You will need to be able to access the cell (as defined by a cell index and a cell-thread pointer) from inside the DPM-based UDF: you can use something like this:

  C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 0) = drag_force;

Or this to calculate and store an average:

#define CUDM_AV_DRAGCOEFF 0
#define CUDM_VISIT_COUNTER 1
  C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_AV_DRAGCOEFF) *= 
    C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_VISIT_COUNTER);
  C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_AV_DRAGCOEFF) += 
    drag_force;
  C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_VISIT_COUNTER) += 1;
  C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_AV_DRAGCOEFF) /= 
    C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), CUDM_VISIT_COUNTER);

(By the way, lots of DPM macros have changed names in recent versions, from P_* to TP_*. Check which version you are using. If you're searching the internet for help, look also for P_CELL etc.)

An alternative approach, which I might prefer, would be to use particle-based user-defined memory -- the so-called "scalars" allocated in the UDF tab of the Discrete Phase Model. These are allocated to each parcel, and you can use them like other particle variables to colour the displayed tracks etc. They are numbered from 0 upwards, like cell-based UDM but completely separate.

  TP_USER_REAL(tp,i) = drag_force;

Good luck!
Ed

[agarhaji]

Hi again.

I tried to use the C_UDMI. I stored drag_force, C_D and it looks like this.

C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 0) = drag_force;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 1) = C_D;

I can now choose in GRAPHICS, User-defined memory and plot the variables

Thank you both !

Have a nice day.

[obscureed]

Great -- well done! To help others, please could you post your working UDF code? (And also tell us which version of Fluent it works in.) Thanks!


I would recommend the particle-based approach as a good alternative.



Ed

[agarhaji]

Im running Ansys fluent 18.2

Compiling steps:

Load Fluent --> User-defined functions --> Load UDF --> Build --> Load Mesh-->SET user-defined memory(in my case, 6.) -->Load UDF --> Write Case --> Close fluent and open again and now the UDF is compiled.

For this UDF, I stored 6 variables.

#include "udf.h"
#include "dpm_types.h"


DEFINE_DPM_DRAG(drag_force,Re,tp)
{

double C_D, drag_force;
double d_sv, L, W, d_Cl, d_CC, Psi;
d_sv = TP_DIAM(tp); /* Sieve Diameter defined as the injected particle diameter */
L = 0.0115*log(d_sv) + 0.0818;
W = -42.365*pow(d_sv, 2) + 1.5702*d_sv - 0.0012;
d_Cl = W; /* Defining the diameter of the inscriped circle as the with of the particle*/
d_CC = L; /* Defining the diameter of the circumscribed circle as the with of the particle*/
Psi = sqrt(d_Cl / d_CC);
/* Calculating the drag coefficient using Yow's correlation */
C_D= ((15.21 + (10.82 / Psi) - (0.14 / pow(Psi, 2))) / Re) + (13.41 - (10.64 / Psi) - (0.016 / pow(Psi, 2))) / sqrt(Re) - 8.82 + (5.70 / Psi) + (0.23) / pow(Psi, 2);
/* Calculating the dragforce as a dimentionless group */
drag_force = (18 * C_D*Re) / 24;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 0) = L;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 1) = W;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 2) = Psi;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 3) = C_D;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 4) = drag_force;
C_UDMI( TP_CELL(tp), TP_CELL_THREAD(tp), 5) = Re;


return (drag_force);

}

After the simulation has converged, the variables can be found under GRAPHICS.


- Agar Haji