Define particle shape

Saidul · January 12, 2015, 05:21

Hello Everyone,
I am working with DPM. Is it possible to define the particle shape by using UDF? By using FLUENT I can easily define the particles size, how can I also introduce particle shape. Can anyone help me?

regards,
Saeed

pakk · January 12, 2015, 08:53

Yes, but it will require a lot of programming/modelling from you.

I don't know which kind of shapes you want, but suppose for simplicity that you want cylinders. Then you can define two Particle scalars, and assign the cylinder height and cylinder radius to these two Particle scalars. Because your particle is now no longer rotationally symmetrical, you need two additional parameters, so you add three Particle scalars, two angles that define the orientation of the axis, and one for the orientation around that axis (or perhaps smarter the angular velocity of rotation around that axis).

That is the easy part.

Then you need a model for how the shape affects the drag, and the change in orientation. This is not implemented, and there is no single model that works best, your best hope is to find a journal article that happens to say what happens to the particles of your shape, and hope that you are able to translate this into a model for Fluent. You need to implement that model in a UDF.

That is still not the hardest part.

The hardest part is to model the wall interactions. If your particles hit a wall, the returning velocity depends on the orientation and rotational velocity of the particle at the moment of impact. And good luck finding a decent model for that...

In other words: yes, in principle it is possible, but no, in practice you don't want this unless you have more than a year to do this, or unless you can accept drastic simplifications, but then it is probably smarter to simplify everything into spheres, and use the standard Fluent implementation.

Saidul · January 13, 2015, 19:37

Thanks for your idea. I will try it later and currently, I am trying to gather some knowledge on it.

Thanks again.

Saeed

fishball · January 17, 2015, 02:28

You could set the DPM model to nonspherical drag law and set the shape factor as chapter 36.3.24 in Fluent User Guide

Saidul · January 18, 2015, 23:09

Quote:

Originally Posted by fishball

You could set the DPM model to nonspherical drag law and set the shape factor as chapter 36.3.24 in Fluent User Guide

Thanks for your information. I will check it later but I have just a curiosity, is it possible to define any shape by changing the shape factor? Suppose i am looking for cylindrical, angular, acicular etc.

Thank you again.

Saeed

fishball · January 18, 2015, 23:51

The shape factor could be estimated by the surface area of a sphere, s having the same volume as the particle, and S* is the actual surface area of the particle.

shape factor=s/S

Simone · January 27, 2015, 12:06

Hello,

I also want to define the particle shape by using udf. Can anybody help me? I don't have any idea how to do that...

Thank you!

fishball · January 27, 2015, 20:25

Quote:

Originally Posted by Simone

Hello,

I also want to define the particle shape by using udf. Can anybody help me? I don't have any idea how to do that...

Thank you!

But why you need to define the particle shape by udf? You could set drag coefficient instead of particle shape in DEFINE_DPM_DRAG.

Simone · January 28, 2015, 04:46

Quote:

Originally Posted by fishball

But why you need to define the particle shape by udf? You could set drag coefficient instead of particle shape in DEFINE_DPM_DRAG.

I need to simulate the body forces, particle-particle interactions and particle-wall interactions for particles with different shapes.

pakk · January 28, 2015, 05:16

Do you know how the body forces, particle-particle interactions and particle-wall interactions depend on the shape of the particle?
Because that is something you have to tell Fluent. It will not be the outcome of the simulation, but is input.

Simone · January 28, 2015, 05:28

Quote:

Originally Posted by pakk

Do you know how the body forces, particle-particle interactions and particle-wall interactions depend on the shape of the particle?
Because that is something you have to tell Fluent. It will not be the outcome of the simulation, but is input.

Yes I know, that I have to tell fluent how body forces and interactions depend on the shape of the particle. And I know how to write an udf for a new body force for example. But first I have to define the shape of the particles. And that's my problem...

pakk · January 28, 2015, 05:42

But the entire point is that you don't define the shape of the particle in Fluent. You only define the effect of that shape.

Simone · January 28, 2015, 05:48

Quote:

Originally Posted by pakk

But the entire point is that you don't define the shape of the particle in Fluent. You only define the effect of that shape.

And there is no possibility to do that?

pakk · January 28, 2015, 06:02

As far as I know, not.

If you use the DPM model, the particles are modeled as point particles, with some corrections. The "standard" corrections make them spherical with a finite size, Fluent has some option included for non-spherical particles at atmospheric pressure, but if you want to do better you have to program it yourself. But whatever you do, these will remain point particles that mimic the shape you want.

With 6DOF modeling, you can put the real geometry of your particle inside. But that is at great computational cost (compared to DPM). I don't think it is feasible to do a simulation of 1000 explicitly modeled particles going through a tube because the mesh would need too many elements, but I never tried that so I am not 100% sure.

Simone · January 28, 2015, 06:31

Quote:

Originally Posted by pakk

As far as I know, not.

If you use the DPM model, the particles are modeled as point particles, with some corrections. The "standard" corrections make them spherical with a finite size, Fluent has some option included for non-spherical particles at atmospheric pressure, but if you want to do better you have to program it yourself. But whatever you do, these will remain point particles that mimic the shape you want.

With 6DOF modeling, you can put the real geometry of your particle inside. But that is at great computational cost (compared to DPM). I don't think it is feasible to do a simulation of 1000 explicitly modeled particles going through a tube because the mesh would need too many elements, but I never tried that so I am not 100% sure.

I think I just can use 6DOF modeling for few particles, but I want to simulate at least 1000. But maybe I can start with 6DOF modeling to simulate the rotation of one particle and then I can think about a model for DPM...
But I never worked with 6DOF modeling. So I have to generate my geometries and the grid for the channel and the particle in DesignModeler and ansys meshing and than I have to write an udf for the behaviour of my particle?

pakk · January 28, 2015, 06:41

I have never used any 6DOF model, I have only read and heard about it, so I can not tell you that...

Simone · January 28, 2015, 06:44

Quote:

Originally Posted by pakk

I have never used any 6DOF model, I have only read and heard about it, so I can not tell you that...

Ok, but anyway thank you for your help!!

Saidul · April 19, 2015, 22:03

Could anyone explain the shape factor for non-spherical particles. As we know, the maximum shape factor value is less than 1. I am not clear about the shape factor value. If I put the shape factor value 0.5 and 0.2, what is the difference between these two cases?

Is it possible to define different shaped particles (rod, cone, cylinder etc) by using shape factor?

Thanks in advance

ghost82 · April 20, 2015, 05:56

Shape factor (PHI), or sphericity was defined by Wadell as:

PHI=(surface area of a sphere having the same volume as the particle)/(surface area of the particle)

Saidul · April 21, 2015, 22:22

Ok, I am a little bit confused about the shape factor value. FLUENT user guide says, the shape factor value must be less than one. But, I found a literature, they said the shape factor value for non-spherical nano particles is larger then 1. They showed, for regular tetrahedral nano particles, the shape factor value is 1.49.

could anyone can help me regarding this issue?
Thanks
Saidul

January 12, 2015, 05:21	Define particle shape	#1
Saidul Member S. Morichika Join Date: Aug 2014 Posts: 62 Rep Power: 12	Hello Everyone, I am working with DPM. Is it possible to define the particle shape by using UDF? By using FLUENT I can easily define the particles size, how can I also introduce particle shape. Can anyone help me? regards, Saeed

January 12, 2015, 08:53		#2
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	Yes, but it will require a lot of programming/modelling from you. I don't know which kind of shapes you want, but suppose for simplicity that you want cylinders. Then you can define two Particle scalars, and assign the cylinder height and cylinder radius to these two Particle scalars. Because your particle is now no longer rotationally symmetrical, you need two additional parameters, so you add three Particle scalars, two angles that define the orientation of the axis, and one for the orientation around that axis (or perhaps smarter the angular velocity of rotation around that axis). That is the easy part. Then you need a model for how the shape affects the drag, and the change in orientation. This is not implemented, and there is no single model that works best, your best hope is to find a journal article that happens to say what happens to the particles of your shape, and hope that you are able to translate this into a model for Fluent. You need to implement that model in a UDF. That is still not the hardest part. The hardest part is to model the wall interactions. If your particles hit a wall, the returning velocity depends on the orientation and rotational velocity of the particle at the moment of impact. And good luck finding a decent model for that... In other words: yes, in principle it is possible, but no, in practice you don't want this unless you have more than a year to do this, or unless you can accept drastic simplifications, but then it is probably smarter to simplify everything into spheres, and use the standard Fluent implementation. Saidul and mimi0201 like this.

January 17, 2015, 02:28	Shape Factor	#4
fishball New Member John Yu Join Date: Dec 2014 Posts: 26 Rep Power: 11	You could set the DPM model to nonspherical drag law and set the shape factor as chapter 36.3.24 in Fluent User Guide Saidul likes this.

April 20, 2015, 05:56		#19
ghost82 Senior Member Rick Join Date: Oct 2010 Posts: 1,016 Rep Power: 27	Shape factor (PHI), or sphericity was defined by Wadell as: PHI=(surface area of a sphere having the same volume as the particle)/(surface area of the particle) __________________ Google is your friend and the same for the search button!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How to define the size of particle in Eulerian	siara817	STAR-CCM+	3	January 15, 2020 15:40
how to define a shape deformation using the grid deformation variables (FSI problem)	momo_sjx	SU2 Shape Design	6	June 7, 2015 23:29
Check particle impaction with User Fortran	Julian K.	CFX	3	January 12, 2012 10:46
How to define Particle with UDF?	sega	Fluent UDF and Scheme Programming	3	January 20, 2010 05:50
how to define particle mass fraction in the subrou	Neser25	CFX	1	October 6, 2005 10:20

January 13, 2015, 19:37		#3
Saidul Member S. Morichika Join Date: Aug 2014 Posts: 62 Rep Power: 12	Thanks for your idea. I will try it later and currently, I am trying to gather some knowledge on it. Thanks again. Saeed

January 18, 2015, 23:51		#6
fishball New Member John Yu Join Date: Dec 2014 Posts: 26 Rep Power: 11	The shape factor could be estimated by the surface area of a sphere, s having the same volume as the particle, and S* is the actual surface area of the particle. shape factor=s/S

January 27, 2015, 12:06		#7
Simone New Member Join Date: Jan 2013 Posts: 7 Rep Power: 13	Hello, I also want to define the particle shape by using udf. Can anybody help me? I don't have any idea how to do that... Thank you!

January 28, 2015, 05:16		#10
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	Do you know how the body forces, particle-particle interactions and particle-wall interactions depend on the shape of the particle? Because that is something you have to tell Fluent. It will not be the outcome of the simulation, but is input.

January 28, 2015, 05:42		#12
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	But the entire point is that you don't define the shape of the particle in Fluent. You only define the effect of that shape.

January 28, 2015, 06:02		#14
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	As far as I know, not. If you use the DPM model, the particles are modeled as point particles, with some corrections. The "standard" corrections make them spherical with a finite size, Fluent has some option included for non-spherical particles at atmospheric pressure, but if you want to do better you have to program it yourself. But whatever you do, these will remain point particles that mimic the shape you want. With 6DOF modeling, you can put the real geometry of your particle inside. But that is at great computational cost (compared to DPM). I don't think it is feasible to do a simulation of 1000 explicitly modeled particles going through a tube because the mesh would need too many elements, but I never tried that so I am not 100% sure.

January 28, 2015, 06:41		#16
pakk Senior Member Join Date: Nov 2013 Posts: 1,965 Rep Power: 27	I have never used any 6DOF model, I have only read and heard about it, so I can not tell you that...

April 19, 2015, 22:03		#18
Saidul Member S. Morichika Join Date: Aug 2014 Posts: 62 Rep Power: 12	Could anyone explain the shape factor for non-spherical particles. As we know, the maximum shape factor value is less than 1. I am not clear about the shape factor value. If I put the shape factor value 0.5 and 0.2, what is the difference between these two cases? Is it possible to define different shaped particles (rod, cone, cylinder etc) by using shape factor? Thanks in advance

April 21, 2015, 22:22		#20
Saidul Member S. Morichika Join Date: Aug 2014 Posts: 62 Rep Power: 12	Ok, I am a little bit confused about the shape factor value. FLUENT user guide says, the shape factor value must be less than one. But, I found a literature, they said the shape factor value for non-spherical nano particles is larger then 1. They showed, for regular tetrahedral nano particles, the shape factor value is 1.49. could anyone can help me regarding this issue? Thanks Saidul