[ch3coohminh]

Hi every body
I have a code about particle size distribution (Rosin-Ramsler)
D = d_tb. (-ln(1- Y ))^(1/n)

PHP Code:

#include "udf.h" 
#include "math.h"
DEFINE_PROPERTY(particle_dia,cell,thread) 
{         
        real n=3; //size distribution constant
        real d_tb=0.0000040; //d_tb is the mean particle size
        real d; // particle diameter
        real y; 
        real v;
        real e=2.71828182845904523536;
        real temp1,temp2;
        real dens;
        v = C_VOLUME(cell,thread);  // get info about volume of cell
        dens = C_R(cell,thread); //get ifo about density of cell
        y=v*dens; //caculate mass fraction of cell
        temp1=(log(-(1-y)))/log(e);
        temp2=exp ((1/n)*log(temp1));
        d=d_tb*temp2;
    

   return d;
  } 


but when i run this code
Error: This is a post-processing only executable.
Error Object: #f
please help me or give a suggest

[ghost82]

Why don't you use the built-in fluent Rosin-Rammler capability?You can specify this distribution without a special udf.

Daniele

[ch3coohminh]

Originally Posted by ghost82

Why don't you use the built-in fluent Rosin-Rammler capability?You can specify this distribution without a special udf.

Daniele

thanks ghost82 for suggest.Can you tell me more details about using built in ROSIN RAMMLER distribution ?
I have not found other way specify this distribution without udf.
thank again so much.

[ghost82]

Originally Posted by ch3coohminh

thanks ghost82 for suggest.Can you tell me more details about using built in ROSIN RAMMLER distribution ?
I have not found other way specify this distribution without udf.
thank again so much.

Sure!
You have to use the dpm (discrete phase model); for example for a steady state simulation, first you solve your problem without the dpm; after you reach the solution, you can enable the dpm model (in the same window where you define turbulence/laminar model, energy equation, etc.); after you enable the dpm, you have to set the max number of time steps (this could be changed if after the particles tracking you have an important amount of "incomplete" particles - incomplete means that particles are still in your domain); you have to define injection(s) of particles (click on injections..) and there you can chose to inject a single, a group, or a surface etc. of particles; you define the type of particles (for example inert), and in the same window you will see the possibility to change the diameter of particles from "constant" to "Rosin-Rammler" or to "Rosin-Rammler-logaritmic".
If you choose Rosin-Rammler you have to input some parameters, based on your particle distribution: see here for the how calculate these parameters:

https://www.sharcnet.ca/Software/Fluent12/html/ug/node692.htm

You may want to enable the stochastic tracking and write the number of stochastic tries: for example, if you set the number of stochastic tries to 2 this means that 2 particles will be tracked in their possible paths, starting from the same point.
You can also choose if your particles interact with the continuous phase or not and after how many iteration the particles interact.

After you define the injection you have to modify the particle material in the material panel.
Then you can modify boundary condition: for example you can set the wall to reflect the particles, and the outlet to let escape the particles.

At the end, to view the particle path you have to set in the graphics and animation window the particle track, choose the release injection point/surface and wait for the result.

Hope this help,

Daniele

[chmurillor]

Hi guys

Do you know any UDF to vary the particle size distribution? I have already included a DPM model and have a solid particles dispersion in air. I want to consider collisions to aggregate and fragment particles during dispersion.

Thanks for suggestions

Carlos

[ch3coohminh]

Originally Posted by chmurillor

Hi guys

Do you know any UDF to vary the particle size distribution? I have already included a DPM model and have a solid particles dispersion in air. I want to consider collisions to aggregate and fragment particles during dispersion.

Thanks for suggestions

Carlos

hi
Please tell more details .

[chmurillor]

The UDF should consider the velocity and position of the particle and the velocity and stresses of the fluid near the particles.

I'd like to know how to invoke these parameters.

Thanks in advance

Best regards

Carlos

[ch3coohminh]

Originally Posted by chmurillor

The UDF should consider the velocity and position of the particle and the velocity and stresses of the fluid near the particles.

I'd like to know how to invoke these parameters.

Thanks in advance

Best regards

Carlos

please contact me by email ch3coohminh@gmail.com
We will disscus about this problem.
Minh Van Nguyen

[souria]

Hey, I know its an old post.
Is anyone can tell me how to introduce a log-normal distribution or a modified Rosin-Rammler in Fluent. Because I had adjusted the distribution law to my own experimental data and I get a modified Rosin-Rammler (and or a log-Normal) which correspond exactly to my data.


Thanks for help

Souria