Uflow usable for pulsejets?

[leo]

I’ am trying to learn how to use Uflow, after getting some strange results, I tried something simple.
I tried 5 different intake pipes, one strait and 4 with different flaring.
The results are almost the same for everyone off them, not what I expected at all.
The conclusion is, Uflow is right and flaring is not needed to get a good inlet flow, or Uflow is not usable for pulsejet design, but only a educational tool like the maker off this software stated.
I go for the last statement, I think the processes in pulsejets are much to complicated to simulate with such a small software package.
I don’t think its good for me to spend more time on Uflow.
Or somebody over here has to convince me otherwise.
I attached the 5 files for analysing.

If you try it with reversed flow you get more mass flow, this means according to Uflow you have to make the flare on the inside of the pulsejet.???
If anybody want to download Uflow here it is.

http://capella.colorado.edu/~laney/uflow.htm

[Lawrence]

I think intake flaring is mostly needed to prevent / reduce the forming eddy's in the intake path, wich restrict maximum airflow.

Perhaps U-flow isn't extended enough to calculate local turbulence like eddy's, or any turbulence at all.. I can't tell, I haven't spend much time with the program yet.

[WebPilot]

leo and Lawrence,

Uflow1d is just that, a learning tool.

Its solver is based on 1d compressible theory, a tuff topic.

Several papers that have been published on pulsejets, use that theory to
attempt to describe the interactions of wave propagation during one or
two cycles. They used a graphical technique called "wave diagrams".

It is a very tedious computation - it may take a day. If you make a mistake
early on, and catch it, you get to do the whole computation (and drawing)
all over again.

UFlow1D can be used as a "check", for your work.

1d compressible flow theory cannot be used to analyze eddies - that is
the realm of 2d and 3d flow requiring faster computational devices.



It will be difficult to analyze a "Chinese", too, with it.

But there is much that can be learned by using Uflow1D, despite its short-
comings.

When I was using it a couple years ago, it didn't have a user manual
expressed in English. Maybe that's no longer true, but I did a translation,
and if you wish, I could share it with you.

You must realize one thing, though, one dimensional unsteady flow is taught
to engineering students at the very earliest at the Senior undergraduate
level. It's no wonder it is so difficult to grasp.

My first wave diagram was that of a gas pulse coursing through a 2
stroke engine's exhaust pipe. This is how they design "extractors".

My fourth or fifth (can't rightly recall, now) was this one (incomplete).



Laters,

-fde

PS Leo, nice anime ...

[Hank]

Hello- The Manual that downloads with Uflow is in Spanish. The Help file is in good English, with enough info to operate the program.

Thanks for this one.

Hank

[larry cottrill]

The "problem" with UFLOW1D is the 1D part - the analysis is one-dimensional. He can't handle multi-dimensional concepts like separation, shear, turbulence, etc. I would guess that the main reason the intake flare doesn't seem to matter much is that the program can't take into account the omnidirectional nature of the flow, since that would be two- or three-dimensional - he just assumes the flow comes straight in. A flare will be seen as a nozzle in one direction and a diffuser in the other.

A good example of this is the inflow from the intake into the dome at the front of the linearized FWE model. The air comes in and brakes almost to a halt inside the dome! (See my posts in the 'UFLOW Progress' thread on the Valveless Forum to see this presented graphically.) UFLOW considers the dome as a very short diffuser, as if the flow would just instantly widen, clinging to the dome surface at the boundary. In fact, this could not possibly happen - the reality will be a jetting of air into the front of the chamber, breaking up into shear and turbulence in abundance.

The trick is just to understand the limitations of the tool (just like you have to do with any other tool you pick up). I think UFLOW1D is excellent at showing general changes in the gas condition, especially if the pipe transitions are kept gradual. The problems are in the detail areas, such as sharp breaks in contour, where no 1-dimensional analysis can describe reality. You just have to learn to understand when you can and can't believe what it shows you.

My old textbook says that you would need to limit a divergent cone angle to no greater than about 12 degrees to completely prevent separation, for air at moderate temperatures (like in the diffuser section of a turbojet)!

L Cottrill

[WebPilot]

The "problem" with UFLOW1D is the 1D part - the analysis is one-dimensional ...

I disagree, here.

Aerovalved pulse jets have been modelled using 1D compressible flow
theory. Comparison with the models' predictions with experiment is good.

The problem with UFlow1D, is, it wasn't written for pulsejet analyses.

It is worth learning how to use this code, though, to improve one's
perceptions of gas dynamics without the rigors of mathematics.

However, GIGO applies.

-fde

[Najm]

Can anyone please post the whole application ( zip form ) here? I cannot access the website which has Uflow, I don't know why, it might be because of rotten internet connection but the page keeps on loading and loading and doesn't stop and the browser window remains blank.

[Al Belli]

Hi Najm,

Here's the Zipped file in the attachment.

Al Belli

[Mike Everman]

Looks like Laney's no longer at Boulder, nor are his link pages. Too bad.

[Najm]

Thanks..

[Hveem]

The "problem" with UFLOW1D is the 1D part - the analysis is one-dimensional ...

I disagree, here.

Aerovalved pulse jets have been modelled using 1D compressible flow
theory. Comparison with the models' predictions with experiment is good.

The problem with UFlow1D, is, it wasn't written for pulsejet analyses.

It is worth learning how to use this code, though, to improve one's
perceptions of gas dynamics without the rigors of mathematics.

However, GIGO applies.

-fde

Hmm.. maybe someone should write a program that does work for pulsejet analysis then ?

[WebPilot]

If you are serious, then :

[Mike Everman]

Hi Forrest! Long time. Made any other putt putts lately?

[WebPilot]

Hey Mike,

I haven't but a friend sent me pictures of an interesting oscillator he built that maybe I can get him to post.

There are never enough hours in the day when it is warm outside.

I've spent much of my time lately doing more practical things like
- working on my MGB (it's in terrible shape cosmetically, but the floor
boards are still sound enough to support an engine block,
- building a speed reducer for a small metal lathe,
- constructing an interface between the tripod and equatorial mount for an
old 3" refracting telescope (it came in a wooden box),
- bringing back to life a 1979-1980 Yamaha XS650 motorcycle and
- replacing the fuel pump (rather replacing the power connector on top of
the tank - someone broke the clips and pushed it back on waiting for it to
fail) on my mother's car.

It's a good feeling when things finally, go your way on a project. It
makes up (sort'a kind'a) for those that don't.

[Mike Everman]

Hey, how far along are you on the speed reducer? Sounds fun. I have a couple of patents on wierd reducers, not variable though. If it's a traction drive, I can send you some "Santotrak" traction grease...