[Ebrahim]

Dear Freinds
I want track particles in a turbulent flow. for modeling the fluctuating velocities, I use a stochastic model which is not in FLUENT, so I should find fluctuating field, and supply these to the Lagrangian particle tracking unit and telling it to add those fluctuations to the mean velocities when doing trajectory computations!
Does anybody know how I can add the fluctuating velocity to the mean velocity?

[BehtashCFD]

Originally Posted by Ebrahim

Dear Freinds
I want track particles in a turbulent flow. for modeling the fluctuating velocities, I use a stochastic model which is not in FLUENT, so I should find fluctuating field, and supply these to the Lagrangian particle tracking unit and telling it to add those fluctuations to the mean velocities when doing trajectory computations!
Does anybody know how I can add the fluctuating velocity to the mean velocity?

Dear Ebrahim, I am facing the same problem, Did you find any way to add velocity fluctuations to the mean velocity used in DPM equations. I appreciate any suggestion.

[Alimohamadi_nasr]

I am not sure about my response, only I guess that maybe it is true.

when you have a turbulent flow it means that you have flow that there are a lot of fluctuations. RANS models define your flow as two parts (mean velocity plus fluctuations) to solve the flow, fluctuations are not solved directly, they are modeled with some equations, such as K-epsilon or K-omega. Therefore, in each cells fluctuation velocities are calculated by using specified model.

In DPM, in usually case for solving Lagrangian equation only mean velocity is used. whenever you choose stochastic model or cloud model, your solver utilizes both mean and calculated fluctuation velocity together. Therefore, when you chose above mentioned model, it is not necessary to add fluctuation velocity, solver will do it by own-self.

[BehtashCFD]

Originally Posted by Alimohamadi_nasr

I am not sure about my response, only I guess that maybe it is true.

when you have a turbulent flow it means that you have flow that there are a lot of fluctuations. RANS models define your flow as two parts (mean velocity plus fluctuations) to solve the flow, fluctuations are not solved directly, they are modeled with some equations, such as K-epsilon or K-omega. Therefore, in each cells fluctuation velocities are calculated by using specified model.

In DPM, in usually case for solving Lagrangian equation only mean velocity is used. whenever you choose stochastic model or cloud model, your solver utilizes both mean and calculated fluctuation velocity together. Therefore, when you chose above mentioned model, it is not necessary to add fluctuation velocity, solver will do it by own-self.

You are right. There are options in FLUENT that particles can be tracked considering the flow velocity fluctuations. Though the stochastic models presented in FLUENT are limited to the Discrete Random Walk (DRW) or Cloud models which are based on the normal random or Gaussian Probability Density Function (PDF). I am looking for a way so that I add turbulence fluctuations to the mean velocity seen by the particles, while they are defined using more sophisticated random functions.

[Ebrahim]

Originally Posted by BehtashCFD

Dear Ebrahim, I am facing the same problem, Did you find any way to add velocity fluctuations to the mean velocity used in DPM equations. I appreciate any suggestion.

Hi Behtash,

There is no special UDF for defining fluctuating velocities. However, you can use other UDFs such as defining Drag Coefficient (DEFINE_DPM_DRAG) or time step (DEFINE_DPM_TIMESTEP). I recommend you to use DEFINE_DPM_DRAG UDF.
When you use such UDF, you can define the drag coefficient in the same way that is defined in FLUENT.

Here is an example. in this UDF, the drag coefficient is obtained from the spherical law (which is one of the forms that is used in FLUENT).

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

DEFINE_DPM_DRAG(particle_drag_force,Re,p)
{
real drag_force;

*************************************
Here you can write neccessary commands for obtaining the fluctuating velocities
and also adding the to the mean flow velocities.
At the end write the command below which defines the drag coefficient the same as FLUENT.
*************************************

drag_force = SphereDragCoeff(p->Re);

return drag_force;
}


Still having questions, do not hesitate to ask.

Bests,
Ebrahim

[wc34071209]

Dear Ebrahim,

Thank you for providing the hint. But one problem is that the fluctuating velocity is a vector and the DEFINE_DPM_DRAG returns a scalar which is the drag coefficient. So it seems this Macro is not something we can use.

Originally Posted by Ebrahim

Hi Behtash,

There is no special UDF for defining fluctuating velocities. However, you can use other UDFs such as defining Drag Coefficient (DEFINE_DPM_DRAG) or time step (DEFINE_DPM_TIMESTEP). I recommend you to use DEFINE_DPM_DRAG UDF.
When you use such UDF, you can define the drag coefficient in the same way that is defined in FLUENT.

Here is an example. in this UDF, the drag coefficient is obtained from the spherical law (which is one of the forms that is used in FLUENT).

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

DEFINE_DPM_DRAG(particle_drag_force,Re,p)
{
real drag_force;

*************************************
Here you can write neccessary commands for obtaining the fluctuating velocities
and also adding the to the mean flow velocities.
At the end write the command below which defines the drag coefficient the same as FLUENT.
*************************************

drag_force = SphereDragCoeff(p->Re);

return drag_force;
}


Still having questions, do not hesitate to ask.

Bests,
Ebrahim

[Ebrahim]

Originally Posted by wc34071209

Dear Ebrahim,

Thank you for providing the hint. But one problem is that the fluctuating velocity is a vector and the DEFINE_DPM_DRAG returns a scalar which is the drag coefficient. So it seems this Macro is not something we can use.

Dear wc34071209,

You use this UDF just to have access to the parameters that you want to change in the particle equation of motion, but what the function returns is still just the drag coefficient. As I explained in the example, you use this function to return spherical drag coefficient, but in the function definition you can add some commands which may / may not have anything to do with the drag coefficient, but changes the parameters that you like.
For example, you can change the particle diameter here or its velocity.

Ebrahim