strange boundary condition in transient sim

• 180 Views
• Last Post 16 January 2018
oberstar posted this 19 December 2017

I had ansys fluent crash and seemingly corrupt my project

I have a eulerian mixing problem with a transient simulation of 2 fluids injecting through concentric tubes at 2 different constant velocities and mixing in a common tube.

My phase 2 material seemingly dissipates across the entire volume at time step 3.

I would expect the phase 2 material (smaller tube to progress down the tube as time progresses).  It seems like the phase 2 inlet boundary condition is suddenly set to 0 at time step 3 or something like that...  all time after 3 appear similar.

(note before the crash I was using a pulsatile velocity on phase 1 via a UDF  but after noticing this volume fraction problem switched it back to a constant)  See attached slide images:

Raef.Kobeissi posted this 20 December 2017

Hello, There are a lot of variables that could affect your result: 1. What is your time step 2. How does your mesh look like? 3. Which model are you using and how does your convergence looks like.

It is difficult to assist without knowing the above parameters.

Regards

Raef Kobeissi

raul.raghav posted this 21 December 2017

As Raef mentioned, can you post a few images of your mesh and the setup you are using for your case. I suspect the time-step to be the issue here.

PS: I investigated the exact same problem 3 years back with OpenFOAM (mixing effectiveness of catheters) !

Rahul

oberstar posted this 08 January 2018

oberstar posted this 08 January 2018

oberstar posted this 08 January 2018

I was using the Eulerian Multiphase, Multi-Fluid VOF Model

Volume Fraction Parameters: Implicit & Sharp Interface modeling

Using only constant Velocity Flows:

Phase 1 Velocity Inlet 2.06m/s

Phase 2 Velcity Inlet 2.79m/s

Both are Magnitude Normal to boundary

Phase 1 Vol Frac is not adjustable @ Phase 1 Velocity Inlet

Phase 2 Vol Frac = 0 for Phase 1 Velocity Inlet

Phase 2 Vol Frac = 1 for Phase 2 Velocity Inlet

Phase 1 Vol Frac = 0 at Phase 2 Velocity Inlet

Outlet defined to have 0 pascal gauge pressure. & phase 2 material is dfined to have 0 backflow volume fraction.

Simulation time is 0.01 sec per step for 60 steps.

I had a version of this running with a time varying phase 1 material via a UDF that used to work but something seems to have changed and that no longer works either.  This is why I went back to the Constant velocity input for phase 1 & 2  to try and debug why things stopped mixing.

Anything else I need to give you in terms of what I configured?

Kindest Regards,

Erick

Raef.Kobeissi posted this 10 January 2018

Hi, From the first observation, I can see that there should be a solid wall between the inlet of phase 1 and phase 2. I wonder how you separated the 2 inlets? Can you please attach the workbench folder .zip and i will have a look at the whole setup.

Regards

Raef Kobeissi

raul.raghav posted this 16 January 2018

Two things that might be causing trouble are:

1. Mesh: I would definitely consider re-meshing the geometry. You should slice the geometry at the region where the mixing begins (when the two pipes becomes one pipe). Slicing it would help you generate a high quality hexa mesh, which is required in this case. A few simple slices would make all the bodies sweepable and would help you a lot in creating a hexa mesh. See attached figures for slicing and meshing your geometry.

2. Timestep: A timestep of 0.01s, according to me, is too high for this case. Is there a reason why you chose this timestep? Go one order of magnitude lower and see if that makes a difference (after the re-meshing!)

Let us know if these points solve your issue (It definitely should in my opinion). Good luck!

Rahul