Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Numerical methods and mathematical models of Elmer
Post Reply
sabah
Posts: 4
Joined: 06 Dec 2017, 21:35
Antispam: Yes

Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Post by sabah »

Hi all,

I'm working on simulating a venous model where the valves open and shut due to pulsatile nanoparticle flow.

To get movement of the valves, I adapted Elmer GUI Tutorial 9: Interaction between fluid flow and elastic obstacle, which uses both the Navier-Stokes and Non-Linear Elasticity solvers.

For the nanoparticle flow, I have tried using particle dynamics. But my results show the added particles as superimposed above the Navier-Stokes flow; rather than being part of it (which is what I want), because ultimately I want to use the model to study particle accumulation.

I've attached my mesh and sif files to run on Elmer non-GUI, and would gratefully appreciate anyone's help/advice on the matter.

Thanks. :)
Attachments
mesh_sif_nanoparticle_help.zip
(91.93 KiB) Downloaded 276 times
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Post by raback »

Hi sabah

This is quite a difficult problem. The challenge is in the interaction of the particles and flow. The true interaction in bidirectional. Let's look at both directions:

If the fluid flow is known then we can follow the particles in that flow. The particles can be either passive tracers or they may also have a momentum equation applied. The models in Elmer currently assume that the viscous drag can only be of the form applicable to flow around a simple objects (such as spheres) surrouneded by a uniform flow. Hence, in this case two approximations would be done. 1) in the boundary layer the velocity gradient over the particle is not accounted 2) the particle-particle interactions are not accounted accurately. Their effect is only considered via their combined effect on the flow.

In this application one would probably be interested what happens with the particles near the walls. Are they attached and under which conditions. The surface phenomena should probably be further developed.

When we know where the particles are they will also have an effect to the macroscopic flow. This would be opposite to the force of fluid acting on the particles. If there are lot of particles in the flow they should have an effect on the viscosity as well. This could be accounted by first computing the fraction of the particles. This can be achieved by sending both active and passive particles and registering at each timestep the volume fraction of active ones. There are some simple examples on such bidirectional coupling but nothing too fancy.

To conclude, the problem is ultimately quite challenging. Probably some qualititive analysis could be within the reach without too much further coding. To truly study the multiscale phenomena accurately would be another story...

-Peter
annier
Posts: 1168
Joined: 27 Aug 2013, 13:51
Antispam: Yes

Re: Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Post by annier »

Sabah,
1. I agree to what Peter has mentioned above and especially the statements on qualitative analysis and multiscale phenomena are noteworthy to consider. Please consider on what he has described.
2. A simple numerical formulation of particles in artery is given at
http://diposit.ub.edu/dspace/bitstream/ ... ns_v02.pdf. You can find it at the link http://hdl.handle.net/2445/66496 provided at page 25 of the manual.


Yours Sincerely,
Anil Kunwar
Anil Kunwar
Faculty of Mechanical Engineering, Silesian University of Technology, Gliwice
sabah
Posts: 4
Joined: 06 Dec 2017, 21:35
Antispam: Yes

Re: Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Post by sabah »

I tried running the Obstructed Artery example (from pg. 25 of the manual - http://hdl.handle.net/2445/66496) as recommended by Anil Kunwar, but it doesn't seem to converge and I cannot get the results shown. I basically converted the mesh and ran the sif file on cmd. I've included an attachment of the files I keep in my directory to run Elmer non-GUI. Am I doing something wrong? Has anyone else tried running this?

Thanks.
Attachments
obstructed_artery.zip
(1.44 MiB) Downloaded 253 times
nicolewells
Posts: 1
Joined: 15 Feb 2018, 08:28
Antispam: Yes

Re: Navier-Stokes, Non-Linear Elasticity and Particle Dynamics

Post by nicolewells »

Hy hai i to have same question and aslo I'm making an application with KINECT SDK 1.8, OpenGL and DirectShow.

My objective is to extract the human with KINECT background removal API and add avi video's frame as a background using DirectShow's IVMR9WindowlessControl9:GetCurrentImage. Then draw about 1000 square textures which have some transparency to make a fire(particle dynamics).

I add video background if the extracted image's pixel's alpha is 0. If the pixel's alpha is 0, I add video frame's pixel in there.

When KINECT didn't detect human, and video frame's pixel fills the background, the fire comes out very well.

The problem is this : When KINECT detected human, and video frame's pixel fills the background except the human area, the fire comes out, but there are few particles microsoft dynamics consisting fire, and sometimes, the other particles which weren't showed up when the human was detect showed up and then disappear in a moment. After the KINECT couldn't detect the human, the hidden particles just show up and it runs very well.

can you guy's help me on this please!

Thanks!
Post Reply