|
[Sponsors] |
November 2, 2008, 03:04 |
How dpm detects the collision with boundary?
|
#1 |
Guest
Posts: n/a
|
Hi everyone,
I try to simulate the discrete phase using dpm, and I am wondering how Fluent can detect the collison between particles and boundary using meshes? There are so many particles and a huge number of meshes in the system. Can anyone give me a detailed explanation? Thanks a lot! |
|
November 2, 2008, 16:40 |
Re: How dpm detects the collision with boundary?
|
#2 |
Guest
Posts: n/a
|
I think FLUENT can do only Eulerian approach and so DPM might be using phase averaged conservation equations (I am not totally sure about this as I never looed at what equations or methodology FLUENT uses for DPM and so this is only a guess). You might want to consider LBM for your simulation.
|
|
November 2, 2008, 18:20 |
Re: How dpm detects the collision with boundary?
|
#3 |
Guest
Posts: n/a
|
FLUENT knows the coordiantes of the centroid and area normal vector of each face on the boundary.
I'm not sure exatly how it detects collision but I would guess it uses the particle trajectory vector and a plane equation for boundary face and checks for intersection. |
|
November 2, 2008, 21:43 |
Re: How dpm detects the collision with boundary?
|
#4 |
Guest
Posts: n/a
|
Thanks.
I think you are probably right. I found this from the fluent website discussing MPM(http://www.fluent.com/about/news/new...i2_fall/a4.htm): "To detect a particle- wall collision, the model identifies the boundary faces (wall surfaces) the particle intersected during the previous time-step,...". But the next questions are: there are so many bounary faces with different orientations and locations in the system, how can Fluent tell which face the particle may pass through? How to implement this contact search efficiently and precisely in Fluent with a large number of particles and meshes? Thanks! |
|
November 2, 2008, 23:05 |
Re: How dpm detects the collision with boundary?
|
#5 |
Guest
Posts: n/a
|
I don't know the answer to your question but I can make an educated guess from my experience writing UDFs.
For any face in the domain, it will have two cells assocaited with it, one on either side of the face. In the UDF manual these are refereed to as cell C0 and C1. For faces on the boundary, cell C1 does not exist. A simple test would be if the particle passes through a face, test if C1 exists. If it doesn't then it is a boundary face. |
|
November 3, 2008, 01:18 |
Re: How dpm detects the collision with boundary?
|
#6 |
Guest
Posts: n/a
|
Thanks again. I am very interested in what you supposed. So could you please tell me where can I find some references talking about the way the face is labelled and also the way to tell a boundary face? I didn't find any related comments by Google.
|
|
November 3, 2008, 04:31 |
Re: How dpm detects the collision with boundary?
|
#7 |
Guest
Posts: n/a
|
you can find that information in FLUENTs UDF manual
|
|
November 4, 2008, 01:33 |
Re: How dpm detects the collision with boundary?
|
#8 |
Guest
Posts: n/a
|
its just an idea,
lets say that you have a particle at position r(x,y,z) and you know time step size delT. Now from this position of particle r you shall be able to get velocity and thus next position for its movement by delT. from this new position r2, if you had a way to find the cell where this new position belongs, it might answer whether the particle just crossed the boundary or not, because if it crossed boundary Fluent will give you an error that cell does not exist. (I think it does if you ask velocity field for the points not insize the mesh). So you could detect whether it crossed boundary or not. |
|
November 4, 2008, 06:34 |
Re: How dpm detects the collision with boundary?
|
#9 |
Guest
Posts: n/a
|
This is defined by the geometry of the droplet relative to the boundary wall. All boundary walls are defined using the mesh and the droplet trajectory are also computed at every time step. So interaction occur at the geomtry represented by the walls and in the dpm model the droplet vanish once it reaches the wall. This is all what I know as a non user of dpm model.
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
About Collision Model | taksue | OpenFOAM Post-Processing | 6 | January 14, 2009 03:10 |
particle collision | Sri | FLUENT | 0 | November 28, 2006 04:07 |
DPM Collision UDF | Dave | FLUENT | 0 | October 23, 2006 05:14 |
DPM drop collision | kiao | FLUENT | 0 | May 31, 2005 05:17 |
droplet collision | kiao | FLUENT | 0 | May 30, 2005 00:01 |