Menu

Discussion Forum

New Discussion
Filter

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Particles disappearing after bounce

Posted Nov 2, 2019, 10:06 a.m. EDT Computational Fluid Dynamics (CFD), Particle Tracing for Fluid Flow Version 5.4 7 Replies

M. Steggink
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

The below post is related to an archived discussion

Hi all,

I tried to model inelastic wall bounces in te particles tracing module, by implementing the formulas in the archived post above. However when I do so, the particles disappear and the warnings I encounter are:

Some particles have been removed from the simulation because of a failure in finding the wall intersection.

Some particles have been removed from the simulation because maximal number of wall interactions within one time step has been reached.

Can anyone tell me why this would happen? I included my model below.

All help would be greatly appreciated!

Attachments:

7 Replies Last Post Jun 20, 2020, 1:58 p.m. EDT
M. Steggink
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Nov 14, 2019, 3:30 a.m. EST
Updated: 5 years ago Nov 14, 2019, 3:30 a.m. EST

Unfortunately, I still encounter the same problem and cannot get help from the support due to a CKL license without regular support. Is there anyone who encountered the same problem, or has an idea where it might arise?

All help is very much welcome!

Unfortunately, I still encounter the same problem and cannot get help from the support due to a CKL license without regular support. Is there anyone who encountered the same problem, or has an idea where it might arise? All help is very much welcome!
Hamid Rezaei University of British Columbia / Renewable Solid Fuel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Feb 15, 2020, 2:58 p.m. EST

Hi,

I also have the same problem. I want to calculate the residence time of solid particles in a rotating drum and COMSOL underestimates the residence time. I believe the reason is that particles bounce on wall and move forward very fast. Have you found a way to reduce the elasticity of particles and have a lower level of bouncing? When I put alpha=0.8 and less, the particles get out of the geometry.

Could you solve this problem? It would be a very big help for me.

-------------------
Hamid Rezaei, Ph.D.
Research Scientist at the University of British Columbia (UBC)
R&D Director at CanDry Technologies Inc.
Adjunct Faculty at New York Institute of Technology (NYIT) – Energy Management Department
Office 604 827 3419 | Cell 778 707 0866
h
Hi, I also have the same problem. I want to calculate the residence time of solid particles in a rotating drum and COMSOL underestimates the residence time. I believe the reason is that particles bounce on wall and move forward very fast. Have you found a way to reduce the elasticity of particles and have a lower level of bouncing? When I put alpha=0.8 and less, the particles get out of the geometry. Could you solve this problem? It would be a very big help for me.
Hamid Rezaei University of British Columbia / Renewable Solid Fuel
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jun 3, 2020, 9:24 p.m. EDT

No, still I am running my model with Bounce B.C. which is alittle exaggeration. I tested a manual correlation with slower bounce velocity on different geometry dimensions and again my particles disappear sometime. Still, I am looking for an answer.

-------------------
Hamid Rezaei, Ph.D.
Research Scientist at the University of British Columbia (UBC)
R&D Director at CanDry Technologies Inc.
Adjunct Faculty at New York Institute of Technology (NYIT) – Energy Management Department
Office 604 827 3419 | Cell 778 707 0866
h
No, still I am running my model with Bounce B.C. which is alittle exaggeration. I tested a manual correlation with slower bounce velocity on different geometry dimensions and again my particles disappear sometime. Still, I am looking for an answer.
Alexandre Arsenault
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jun 4, 2020, 9:19 a.m. EDT

Hi all,

I am not an expert in the particle tracing module, but I believe that Comsol removes particles when too many wall collisions occur within a time step in order to reduce computation times. Is it possible that when using the inelastic bounce, the particles end up "stopped" at the wall, and therefore bounce indefinitely?

It is possible to increase the maximum number of wall interactions per time step under "particle tracing"->"advanced settings", but this is not a very ideal solution. Ideally, it would be nice to be able to freeze particles after a certain number of wall interactions per time step, but I do not know if this is possible.

Cheers,

Alex

Hi all, I am not an expert in the particle tracing module, but I believe that Comsol removes particles when too many wall collisions occur within a time step in order to reduce computation times. Is it possible that when using the inelastic bounce, the particles end up "stopped" at the wall, and therefore bounce indefinitely? It is possible to increase the maximum number of wall interactions per time step under "particle tracing"->"advanced settings", but this is not a very ideal solution. Ideally, it would be nice to be able to freeze particles after a certain number of wall interactions per time step, but I do not know if this is possible. Cheers, Alex
Christopher Boucher COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jun 4, 2020, 12:04 p.m. EDT
Updated: 5 years ago Jun 4, 2020, 12:04 p.m. EDT

Hi,

As Alexandre writes above, the fallback behavior is for particles to disappear if they bounce too many times in a single time step taken by the solver.

Here are some other situations in which the fallback Disappear behavior may appear:

  • The reflected particle velocity points into the boundary instead of away from it. If it's an interior boundary, the particle can just end up on the opposite side, but for an exterior boundary the particle will just disappear. This can happen with the General Reflection condition if the user-defined expression points into the wall.
  • The particle is reflected off a concave (curved) surface and the reflected particle velocity is nearly tangent to the surface. This is more likely to make particles disappear if the curved surface is coarsely meshed.
  • A Drag Force is also applied to the particles, and the time steps are too large. The characteristic velocity response time for particles in a fluid is (particle density X particle diameter squared) / (18 X fluid dynamic viscosity). Generally you don't want this quantity to be too much smaller than the time steps taken by the solver.

Best,

Chris

Hi, As Alexandre writes above, the fallback behavior is for particles to disappear if they bounce too many times in a single time step taken by the solver. Here are some other situations in which the fallback **Disappear** behavior may appear: * The reflected particle velocity points into the boundary instead of away from it. If it's an interior boundary, the particle can just end up on the opposite side, but for an exterior boundary the particle will just disappear. This can happen with the **General Reflection** condition if the user-defined expression points into the wall. * The particle is reflected off a concave (curved) surface and the reflected particle velocity is nearly tangent to the surface. This is more likely to make particles disappear if the curved surface is coarsely meshed. * A **Drag Force** is also applied to the particles, and the time steps are too large. The characteristic velocity response time for particles in a fluid is (particle density X particle diameter squared) / (18 X fluid dynamic viscosity). Generally you don't want this quantity to be too much smaller than the time steps taken by the solver. Best, Chris
Alexandre Arsenault
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Jun 4, 2020, 12:35 p.m. EDT

Hi Chris,

Thank you for the clarification. Is it possible to change the particle's behaviour after the maximum number of wall interactions is achieved instead of having them disappear?

Cheers,

Alex

Hi Chris, Thank you for the clarification. Is it possible to change the particle's behaviour after the maximum number of wall interactions is achieved instead of having them disappear? Cheers, Alex
john grace
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 years ago Jun 20, 2020, 1:58 p.m. EDT

Hi,

As Alexandre writes above, the fallback behavior is for particles to disappear if they bounce too many times in a single time step taken by the solver.

Here are some other situations in which the fallback Disappear behavior may appear:

  • The reflected particle velocity points into the boundary instead of away from it. If it's an interior boundary, the particle can just end up on the opposite side, but for an exterior boundary the particle will just disappear. This can happen with the General Reflection condition if the user-defined expression points into the wall.
  • The particle is reflected off a concave (curved) surface and the reflected particle velocity is nearly tangent to the surface. This is more likely to make particles disappear if the curved surface is coarsely meshed.
  • A Drag Force is also applied to the particles, and the time steps are too large. The characteristic velocity response time for particles in a fluid is (particle density X particle diameter squared) / (18 X fluid dynamic viscosity). Generally you don't want this quantity to be too much smaller than the time steps taken by the solver.

Best,

Chris

Hello my friend Thank you very much for your responsibility and your response. Let me first explain my geometry to you. First, suppose you have an oblique cylinder channel. Particles move from the top of the channel under gravitational and drag force. Due to the gravitational force, the particles hit the lower surface of the channel. I want to use a boundary condition with which I can describe the motion of particles after they hit a wall. Is the bounce boundry condition correct in this case or not? If I use the correct boundary condition, thank you for giving me a solution to solve this problem. Regarding the first part of your orders, do you mean that if the bounce boundry condition causes warning, you can use General Reflection instead of bounce? In the case of the second part, given the geometry of the problem described above, my geometry is curved (cylindrical), but there is still an error when the mesh is smaller. In the case of the third part, the said parameter for my physics is about 0.027. What do you idea about the right time step? with all respect

>Hi, > >As Alexandre writes above, the fallback behavior is for particles to disappear if they bounce too many times in a single time step taken by the solver. > >Here are some other situations in which the fallback **Disappear** behavior may appear: > >* The reflected particle velocity points into the boundary instead of away from it. If it's an interior boundary, the particle can just end up on the opposite side, but for an exterior boundary the particle will just disappear. This can happen with the **General Reflection** condition if the user-defined expression points into the wall. >* The particle is reflected off a concave (curved) surface and the reflected particle velocity is nearly tangent to the surface. This is more likely to make particles disappear if the curved surface is coarsely meshed. >* A **Drag Force** is also applied to the particles, and the time steps are too large. The characteristic velocity response time for particles in a fluid is (particle density X particle diameter squared) / (18 X fluid dynamic viscosity). Generally you don't want this quantity to be too much smaller than the time steps taken by the solver. > >Best, > >Chris Hello my friend Thank you very much for your responsibility and your response. Let me first explain my geometry to you. First, suppose you have an oblique cylinder channel. Particles move from the top of the channel under gravitational and drag force. Due to the gravitational force, the particles hit the lower surface of the channel. I want to use a boundary condition with which I can describe the motion of particles after they hit a wall. Is the bounce boundry condition correct in this case or not? If I use the correct boundary condition, thank you for giving me a solution to solve this problem. Regarding the first part of your orders, do you mean that if the bounce boundry condition causes warning, you can use General Reflection instead of bounce? In the case of the second part, given the geometry of the problem described above, my geometry is curved (cylindrical), but there is still an error when the mesh is smaller. In the case of the third part, the said parameter for my physics is about 0.027. What do you idea about the right time step? with all respect

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content