CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Dynamic Mesh on Pintle type injector. (http://www.cfd-online.com/Forums/fluent/68699-dynamic-mesh-pintle-type-injector.html)

herntan September 29, 2009 09:02

Dynamic Mesh on Pintle type injector. (Picture and Video Attached)
 
Hi ,

I currently working on a pintle type injector with the shape as below
http://lh3.ggpht.com/_eF1ePnj1QJY/Ss...HQ/Pintle1.jpg
http://lh5.ggpht.com/_eF1ePnj1QJY/Ss...ko/Pintle2.jpg
http://lh4.ggpht.com/_eF1ePnj1QJY/Ss...2Y/Pintle3.jpg

I had tried to create a dynamic mesh model with the Pintle moving down.
The fuel flow rate should getting higher and higher together with the pintle move down.
In the early few steps, the dynamic mesh was working well, i guess its spring smoothing working,
but after a few time step, the mesh fail and get negative volume.
what i can guess is the remeshing and layering is failed.

I not really familiar with the dynamic mesh setting.
I have no idea what mesh should i prepare for this and what dynamic meshing i should be using.

Hope there is someone can tell me how this dynamic mesh can be done?


Thanks a lot. :)

herntan September 29, 2009 11:56

Video of simulation.
 
Here is the simulation video.
I have small stroke of pintle movement simulation done.
But once i move the pintle stroke higher, the simulation fail with negative volume.

http://www.youtube.com/watch?v=w-mi5GZgalA

-mAx- September 30, 2009 01:34

I am not familiar with smoothing, but I assume it is only available for small displacement.
Smoothing is just a distorsion of your mesh, but there is no remeshing.
So you may enable the remeshing option.
But in you case I would recommand Layering (add/suppress cell layer)

herntan September 30, 2009 02:51

I think i will try to do the layering, thanks for the suggestion.

The simulation should have the injector fully shut off by the pintle head initially.
I not able to fully shut it off and it will always leaking initially as in the video.
http://www.youtube.com/watch?v=w-mi5GZgalA

any idea what technique i can do to make it fully shut off by dynamic mesh?

-mAx- September 30, 2009 14:43

I don't see anything on the video (opening)
Describe the opening you want to modellise (sketch)

herntan September 30, 2009 21:00

Max,

Thanks for the reply.

Here is the picture of minimum gap that i wish to close.
Due to meshing issue, i was never able to close up this gap so there is no fuel flow out before the pintle move down.
If u see the video again, u will see there is fuel flow out early in the video without pintle moving down.
http://www.cfd-online.com/Forums/mem...50-pintle4.jpg

Here is the video link.
http://www.youtube.com/watch?v=w-mi5GZgalA

-mAx- October 1, 2009 01:47

you cannot close properly your gap.
Else you become negative volume
If the gap is really axisymmetric, you can use layering and supress cells (gap smaller and smaller)
Play with parameters and cell size to resolve the close of your gap.
Also if your time step is too big, you won't resolve it properly.
But physical closing is (in my sense) impossible because of meshing topoloy

herntan October 1, 2009 02:01

Max,

Thanks alot for clearing my doubt.
I still wondering if i am stupid for not able to close the gap.
The gap is axisymmetric,but its a very small gap ~0.1mm.
I will have the simulation start by deactivate the injector zone and once i activate it will together with the pintle movement.
This should give a nice mass flow rate graph.

Again, Thanks for the reply.

-mAx- October 1, 2009 02:44

you can close the gap, but you cannot have contact between your seat and your pintle.

You can exercise yourself by solving it in 2d axy-symmetric, you will fix your problem easier and quicker
;)

herntan October 1, 2009 02:54

You mean it is possible to have no leak ?
I still wondering if i able to create a gap which Fluent will take it as contact.

I need this injector in 3D. I got a non symmetric chamber on my future simulation.

I am planning on create mesh for the cone shape cavity which the pintle sitting on.
Normally for layering mesh, i will use the extrution with mesh in gambit.
It works fine for engine piston movement.

-mAx- October 1, 2009 03:07

you can define a no-leak event (define/dynamic mesh/events), but it won't be caused from solid-contact.
you can modify your inlet BC at this special time, so that there is no massflow.
But I never did that, so maybe I am wrong...

herntan October 1, 2009 04:27

The event at dynamic mesh is a very funny thing i found in Fluent.
I was never able to change BC in the event.
The best i can do is copy BC in the event.
I have no idea how to go around with this but to create a useless BC at somewhere else for it to copy.

The choices in the event is very limited.
Yes, i can change the pressure inlet become Wall. But i cannot change the Wall become pressure inlet.
Do you think there is any technique to switch between pressure-inlet and wall ?

-mAx- October 1, 2009 04:51

As I said I have no experience with events, so I cannot answer..
I would try to set the MDM with 2d-axi.
If you set the layering properly, you can iterate until there is only one layer in the gap. Only after you will have mesh issue.
But if you have 100 layers in your gap, the you can close your gap with a 99.e-3 mm - stroke.
If the results are ok, get inspired from it, and switch to 3d

herntan October 1, 2009 05:03

OK. Thanks Max.

I will post the video here if i successfully done it.

Anyhow, i still wish someone can teach me how to use the event in dynamic mesh to switch from wall to pressure inlet.

yubaibai88 April 3, 2010 01:22

Quote:

Originally Posted by herntan (Post 230856)
Here is the simulation video.
I have small stroke of pintle movement simulation done.
But once i move the pintle stroke higher, the simulation fail with negative volume.

http://www.youtube.com/watch?v=w-mi5GZgalA

hello herntan:
youtube is unable to log in in my country,would you sent these video to my e-mail.
thank for your help

rahgoshafan January 4, 2012 04:31

Hi all
I need help for simulation a pintle injector .
can send me project file?

thanks


All times are GMT -4. The time now is 02:33.