CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Particle Tracking (http://www.cfd-online.com/Forums/fluent/48693-particle-tracking.html)

Corentin July 7, 2008 03:24

Particle Tracking
 
Hi!

I am using Fluent to calculate trajectories of particles in a plasma flow.

I have been doing it in steady flow but I am now doing it in unsteady flow.

My problem is : in unsteady the trajectories are COMPLETELY different than the ones I have seen in steady, while the plasma flow is more or less the same.

Its seems particles have gained a lot of inertia because they barely not taken in the plasma jet.

I would agree if there were small differences, but there are really big differences.

I guess I made a mistake somewhere.

But I don't have a clue.

What could I made possibly wrong? I just changed for unsteady in the models pannel. I activated unsteady tracking for particles too. The flow is now unsteady but isn't really different from the steady one.

I checked by putting it in steady again and looked at the particles, they are taken as before.

Is there something I could have done wrong?

Thanks!

Joshua Strodtbeck July 14, 2008 08:18

Re: Particle Tracking
 
If you just changed to unsteady in the models panel, then it sounds like you're still using a RANS turbulence model. Your k-e and S-A models are based on equations in which the time derivative disappears due to integration in time, so LES and DES are your only real options in unsteady flow. Yes, I know Fluent lets you use k-e, k-w, and S-A in unsteady flows, but if you do the math carefully, you'll see this is something of an oversight.

Second, your "steady" model isn't really computing steady-state flow. Turbulent flows have no steady-state solution. What it's attempting to do is compute a time-averaged flow field. The actual turbulent phenomena can differ significantly from the time-averaged picture.

Also, make sure your time step is small enough. LES will crash if it's too big.

Corentin July 14, 2008 08:31

Re: Particle Tracking
 
Thank you.

I know that k-e is doing time averaging and I know it is not very relevant in unsteady flow, I even opened a discussion on that point some time ago.

But I am supposed to use k-e because this is a k-e modifyed for plasma flow, which has been validated, and I don't have any LES model that would be relevant for plasma.

Anyway, I have been asked to use k-e, I am not supposed to change it.

Do you think that the very big differences I am seeing are caused by the differences between steady and unsteady k-e?

The plasma I used for steady particle calculations was just taken from an unsteady k-e simulation that I stopped before injecting particles... so, I still don't understand when I am seeing such different trajectories.



All times are GMT -4. The time now is 04:56.