CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Particle Track (http://www.cfd-online.com/Forums/fluent/29536-particle-track.html)

Zahir March 19, 2002 14:52

Particle Track
 
HI I have a cylindrical combustion chamber set-up in fluent, but instead of modelling the full chamber i have chosen a 60 degree segment, with periodic boundaries, the injector creates a swirling flow and i have a decent solution, so far,

but I am hoping if fluent can use the boundaries to do a particle track in the full 360 degrees.??

would this require creating a symmetry of my current segment six times, i hope not cause i'm already struggling with computing memory.

thankx

anna March 19, 2002 16:59

Re: Particle Track
 
just go to display/views, and under periodic type in 6, click on Apply, and you will get 3d display of your geometry.

regards, anna

Yvonne March 20, 2002 05:33

Re: Particle Track
 
Hi Zahir and Anna!

You can do this with the periodic view, but it won't show you exactly what you want. Have you tried it? You will see, that the particle, leaving the periodic side will appear on the other periodic side again, which makes sense, because its what you want. But when you just take the display 6 times, you will have THAT DISPLAY 6 times, you won't get the particle track correctly! This only helps for showing the whole geometry, not the particle tracks and I don't think there is any possibility to get it right without solving the whole geometry. But thats a little bit senseless, just to get nice pictures:)

Anyhow, goog luck, Yvonne

Phil March 21, 2002 04:48

Re: Particle Track
 
I am confused by your answer, if the flow is truely periodic then why would you not expect the whole flow to look like the array of the periodic segment. Isn't this the definition of periodic? I'd be interested in your reply as I am running a periodic analysis presently .

Cheers

Phil

Yvonne March 21, 2002 05:07

Re: Particle Track
 
hi Phil!

Yes, i guess my answer was a liitle bit confusing. It's easier to undrstand, if you try it. if you are solving a continous phase under periodic condition, everything is fine. You 'll get correct results. The problem here is just, that the periodic DISPLAY of a particle track doesn't make real sense. If you display a particle track in a periodic calculated segment, the particle, going out at one periodic side, comes in again at the other side. If you use now the periodic DISPLAY, the particle still comes out at the other side and won't be displayed in the next segment, as it is in reality. so using the periodic feature to have less capacity problems is for sure meaningful. And also particle tracking makes sense. But you won't get the real particle track by just DISPLAYING priodicly. I hope that was clearer. Sorry for the confusion! Yvonne

Phil March 21, 2002 05:40

Re: Particle Track
 
I think I am probably confusing particle tracks and path lines, Is a path line equivalent to a particle track with an injection that has the same properties as the continuous phase? I have a problem with swirl, when I display the path lines in one segment then using periodic repeats of the display the path lines are as expected...helixes.

Yvonne March 21, 2002 08:22

Re: Particle Track
 
hi Phil!

Hmm, that's interesting with the pathlines, perhaps I should try that. I used the particle tracking, where I made an injection with a socalled tracer (same properties as the continous phase) and then I displayed the particle track by choosing an ID-number of one particle stream. But I haven't tried it with the pathlines. Thanks for the tip! Bye, Yvonne:)

Yvonne March 21, 2002 10:33

doesn't work
 
Hi again! I tried it but for me it's not suitable. I get pathlines from every cell in one surface, which are over 4000 in my case and like that I don't see anything! Can you just dislay one single pathline? Yvonne

Phil March 21, 2002 10:42

Re: doesn't work
 
you cant display one single pathline but you can use skip in the pathlines menu, if the skip is set to 0 then a pathline is drawn from every cell, if it is set to 1 then it will draw a pathline from every other cell etc, I think the maximum you can set it to is 200. hope this helps

Yvonne March 21, 2002 11:19

Re: doesn't work
 
Hi Phil! I'm still not convinced. I tried it and you are right, the maximum number to skip is 200. So for displaying pathlines of a continous phase its ok with the repeating display but not for the particle tracks, because you won't see the right track or better said: you see the right track but too often (that often as you repeated it). it's the same with the pathlines but there it doesn't matter because , I guess, they just show the principle way of the streamlines of continous phase, don't they? It's still confusing, especially because I have to write in English:) Have a nice day, Yvonne:)

Phil March 21, 2002 11:50

Re: doesn't work
 
Yvonne, I have just been having a play with injections and particle tracks, I think I can get the display using particle tracks with the injection set as a surface injection. Then in the particle track menu picking the track particle stream button and choosing the particle id to track. I am using a 3rd model so it looks like it is tracking 3. I get helixes as I am expecting and colouring by residence time suggests that the particle is being tracked out of on side of the boundary and back through the other boundary multiple times until the particle leaves the model. I can send you a screen shot if that would help. Your english is very good, I certainly wouldn't be ab;e to discuss this in german!

Phil

Yvonne March 22, 2002 05:01

Re: doesn't work
 
Good morning, Phil:) I wanted to reply yesterday, but I guess the server ran down. First:thanks for the 'English compliment':)

Ok, the never ending story. I did exactly the same as you. In my case I have a 120 segment, which I repeated 2 times. Of course, we can see helixes, but in my case, I get 3 helixes, exactly the same. You are right, there is one particle track which is shown right, but additionally two, which we don't want. If you look at it in time steps (with one time step I mean the time, the particle needs to go through one segment), it's like that: the particle starts in the first segment, comes to the periodic side and appears again in the second step in the same segment on the other periodic side. If you repeat the display, this first particle track appears in the other 2 segments, too. And of course also the next steps will appear. Now, the second step of the second segment fits to the first step of the first segment and fits to the third step of the third segment, so we get a continuous helix. But we get it 2 times too often. Or do you just get one helix and I'm again writing very confusing stuff? If it's like that, please send me your one helix! Till the next reply:) Yvonne

Phil March 22, 2002 06:45

Re: doesn't work
 
No I get three helixes when selecting only the one particle ID. However I do not think that the particle track is wrong. Looking at one segment (120degrees) it does appear that the particle leaves one side of the periodic boundary and appears instantly on the other side of the periodic boundary (and from the way that the boundary is defined i am sure that is what fluent is calculating). This is also exactly the same result as you would expect to see if you were looking at 3 particle tracks each 120 degrees apart but only looking at a 120 degree segment.

I'll have to admit defeat about displaying only one helix, unless you just want a 2d picture in which case you could always use a graphics program to edit out the extra particle tracks.


Yvonne March 22, 2002 09:00

Re: doesn't work
 
Hi again! Is it really the same result as you would expect, solving the whole geometry? I'm not too sure. On the one hand side it sounds obvious but on the other...I mean, you use the stochastic model (or, I used it), so I guess, we would get different tracks for each particle. But I won't try ...it would take too long. Bye, Yvonne:)

Zahir March 22, 2002 11:23

Re: doesn't work
 
phil and yvonne it seems both of you are tring to create the same helix i am, i am going to try the pathline technique that phil suggested ie. a pathline with an injection and see if it works, but creating this image is not a neccesary part of my report, but i didn't think it would be this complex. i'll keep you posted.... thanks

Phil March 25, 2002 07:26

Re: doesn't work
 
I had stochastic tracking turned off, I think you are right about the tracks being different if the stocahstic model is used. If i can find time I might try and solve a very simple geometry using periodic boundaries and a full model to see what the differences would be. Unfortunately time is at a bit of a premium at the minute so it won't be soon!

david March 31, 2002 01:29

Re: doesn't work
 
"Is it really the same result as you would expect, solving the whole geometry?"

NO


All times are GMT -4. The time now is 20:57.