Hi,

I am trying a simple simulation of multiphase through an elbow, and the simulation concludes with convergence attained. However, when I view the streamlines for the particles they travel through the pipe into the elbow wall and stay there. The particles should be bouncing off the wall I thought. Any ideas? Here are my simulation paramaters. 0.01% volume solid, 0.99% fluid (air) inlet 5m/s, eulerian-eularian approach particle density 1500 kg/m3 mostly default settings otherwise.

Hi,Tim. I'm interested in your problem,but could you please first tell me how I can make a 2D simulation in CFX? Since so far as I know, CFX has only 3D solver,how could you solve a 2D problem with it？

To make a 2D problem in Ansys Workbench you need to use extruded 2D meshing with one element thickness in the meshing stage. Then when you are setting up your simulation you use a symmetry boundary condition for the two surfaces. Probably best if you use help to better explain it. I think there is also a tutorial that uses 2D meshing.

Tim.

Hi,

The 2D question has been asked a million times before.....

http://www.cfd-online.com/Wiki/Ansys..._simulation.3F

Also look in the documentation.

I hope ANSYS sees the need for a true 2D solver in the next generation CFD code.

Glenn Horrocks

Hi,

The eularian approach cannot model particles bouncing off walls. If you want to include this you need to go to a lagrangian particle tracking model.

Glenn Horrocks

Hi Glenn,

Ok thanks for your response, I was unaware that Eulerian could not model particles bouncing off walls. So it can be said that CFX software cannot model large volume fractions (50%) of particles that interact and bounce off walls, since Langrangian can only be used for dilute multiphases of solid particles. Software using DEM (e.g. EDEM) would be more necessary then correct?

Tim.

Glenn,

Another quick question, in a Eulerian simulation when the particles hit the wall what happens? Do they exit the domain, remain in place where they hit or continue along the wall with the same velocity?

Tim.

Hi,

In V11 particles can either stick or bounce. I can't remeber but I think you can set the normal and tangential coefficient of restitution separately so if that is the case you can make them slide along walls too.

In V12 there will be more options I believe.

Glenn

Hi,

Correct - the Lagrangian model cannot model particle to particle interactions. I think V12 has some interaction stuff and some other goodies for lagrangian models. But if you want to do it now you need to investigate EDEM or other DEM software.

Glenn Horrocks

Glenn,

I am talking about a Eulerian simulation where the particles cannot bounce off a wall. If I run a simulation in Eulerian and follow a streamline of the particles they run into a wall and stay there - numerically what does the solver do with them? Are they stuck to the wall permanently or do they exit the domain?

Tim.

Hi,

I am not sure in V11, but again in V12 (coming soon hopefully!) I think there is some additional physics coming to do wall films. It is aimed at fuel injection in engines but it may have applications elsewhere.

I would talk to support for more details.

Glenn Horrocks