Valveless pulsejet-computational fluid dynamics

Moderator: Mike Everman

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Fri Apr 28, 2017 5:38 pm

Mike Everman wrote:Oh, that one is fun to watch! I would love to see the pressure animation as well. Does it not make sense that the velocity scale should be -300 to +300m/sec? Loses that "hot-cold" spectrum, but freezing it would allow some better interpretation.
Will surely post the pressure plots in my next simulation. -300m/s i dont think so, the velocity reduces to about 0.06m/s near outlet, i took this as ref and setup the velocity limits in scene plot.

Mike Everman
Posts: 5007
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by Mike Everman » Fri Apr 28, 2017 5:46 pm

The velocity has a direction vector left and right. Maybe a silly thing I'm wondering, as we see the propagation of the velocity changes, which does indicate a change in direction.
I'm trying to wrap my mind around whether I can be fooled by interpreting the propagation of the velocity change with gas movement direction. I mean, is it possible that a gas movement could be going opposite to the visual movement of the colors?
Mike Often wrong, never unsure.
__________________________

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Fri Apr 28, 2017 6:27 pm

Actually there will be a velocity motion in opposite direction during suction phase from exhaust and intake side but velocity does not show in negative value (i.e positive velocity in reverse direction). In this video the motion of colours represents the movement of mass either in froward or reverse direction.

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Sun Apr 30, 2017 6:17 am

Another valveless type numerical simulation but this time its single intake chinese type. Its really interesting to see the things happening inside the PJ.

https://youtu.be/FgFmT-_KZfA
Attachments
2017-04-30 11_47_16-New notification.jpg

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Thu May 11, 2017 2:21 pm

A just for fun simulation.

Video link:
https://youtu.be/AlN7lFJ826o
Attachments
2017-05-11 19_50_36-New notification.jpg

Mike Everman
Posts: 5007
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by Mike Everman » Thu May 11, 2017 2:47 pm

Ha! Nice. Oh yeah, I need to dig up jet dimensions for you!
Mike Often wrong, never unsure.
__________________________

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Mon May 15, 2017 1:46 pm

Another just for fun numerical simulation. This was fun creating.
https://youtu.be/VdnT-rudMb8

Mike Everman
Posts: 5007
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by Mike Everman » Wed Jun 21, 2017 1:35 am

Ha, nice. Sure does seem that the intake and exhaust alternate. I truly think that u-bent motors benefit from this effect.
Mike Often wrong, never unsure.
__________________________

Mark
Posts: 10931
Joined: Sat Oct 11, 2003 10:14 pm

Re: Valveless pulsejet-computational fluid dynamics

Post by Mark » Wed Jun 21, 2017 2:44 pm

I didn't see any ear muffs on that stick figurine. The other day reading some Lockwood type literature, it said the sharp/torturous 180 u-bends lose a bit more thrust than the more gradual 180 bends. My snorkelers don't like u-bends but they will tolerate them, they're more fragile.
Presentation is Everything

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Mon Jul 03, 2017 6:09 pm

Hi all,

Recently I have been thinking to write a CFD code specifically to simulate pulsejet engines. I started the work two weeks ago and now I have reached a starting point with fully compressible coupled Navier Stokes equation (2D and 3D analysis). Currently tweaking some of the things to match our requirements, I will release the GUI soon in public so that people can use for their research work.

This code can be run using either MATLAB or Octave.

Some images attached with this message.
Attachments
v1000.jpg
Lid driven cavity
pressure.jpg
Flow past rectangle
bacward_step.jpg
Backward facing step
flow _pipe.jpg
Flow through pipe

Mike Everman
Posts: 5007
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by Mike Everman » Tue Aug 29, 2017 4:42 am

Hey Ganesh, how is the project going?
Mike Often wrong, never unsure.
__________________________

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Sat Sep 02, 2017 10:58 am

Mike Everman wrote:Hey Ganesh, how is the project going?
Hi Mike

Got some updates, In my previous post you can see some colourful images it seems the Lax-Friedrich scheme is too dissipative. So now I have changed the scheme to Maccormack scheme (predictor-Corrector method), the results seems to be reasonable. I have uploaded aa video here of Flow past cylinder case for coarse grid,

https://www.youtube.com/watch?v=h2tYoRe_ByI

The above video is a fully compressible NS laminar regime.

Recent code: Compressible URANS(Unsteady Reynolds averaged Navier Stokes) equations with K-Omega (Wilcox model) turbulence model. Explicit method seems to diverge so currently trying to implement Algebraic multigrid method (Iteratve method). Regarding the combustion model I am going for Non premixed turb combustion (Once AMG is implemented next is combustion).

Regards
Ganesh

Mike Everman
Posts: 5007
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by Mike Everman » Sat Sep 02, 2017 7:01 pm

Great!
Mike Often wrong, never unsure.
__________________________

ganuganu
Posts: 346
Joined: Tue Dec 14, 2010 7:31 am
Antipspambot question: 125
Location: india
Contact:

Re: Valveless pulsejet-computational fluid dynamics

Post by ganuganu » Sun Nov 19, 2017 6:46 am

Got some updates on the code creation,

1. 2d RANS successfully done in MATLAB. STill not yet reached to combustion coding.
2. Created a basic GUI for users to enter values and visualize answers.

As of now I have written 1d, 2d inviscid and laminar flow in explicit form. Trying to build code for implicit so that we can go for larger timesteps and provide some stability to solution. Please find attached images of the GUI created. Once I have written implicit form soon to release the GUI, you can use and provide your feedback regarding GUI usage or any bug issue. As of now I am limiting my code to 2D structured, soon to implement unstructured 2D and 3D.

The reason I am taking littel longer time is that first I will write in MATLAB lang in which I am very comfortable and later on i will write the same in C++.
Attachments
qt.png
Main GUI
qt3.png
Boundary condition & Initial
qt2.png
Results (Graph)
qt1.png
Mesh creation/Solver settings

zono
Posts: 78
Joined: Wed May 28, 2008 1:03 pm

Re: Valveless pulsejet-computational fluid dynamics

Post by zono » Sun Dec 17, 2017 12:38 am

Dear Ganesh,

nice work here!

Some time ago I did some CFD on valveless pulsejets using Comsol Multiphysics. There I used compressible NS formulation and an additional heat-source term as pseudo-combustion. Later I will send some images of that.

Your attempt to create a free / opensource code for octave and Matlab is a honorful work !

Hence I am medium-experienced in OpenFOAM I will try do do the same work with this great tool. Do you know OpenFOAM ?

Best Regards,

zono
___________________________
zono - the thruster buster
--------------------------------------

Post Reply