UFLOW1D experiments with the "Squirrel"

[toakreon]

As I said elsewhere, I've put my very rough-and-ready "design" for the "Squirrel" valveless pulse jet into UFLOW1D and had a look to see what happens.

A LOT happens, actually! But as I am unsure about what really SHOULD happen, I am equally unsure about whether what I'm seeing is "good" or "bad". What I've done is taken a series of "snapshots" of the UFLOW1D results at what I hope are "interesting points" and I'll post them here to see what others think of them ... :)

I also had to guess my initial conditions for the simulation. I don't know how valid these are (comments appreciated) but I tried:-

INLET - divided into two equal length sections. Gas temperature at the "open air" end set to 600K. Gas temperature at the "combustion chamber end" set to 1300K. Both had an initial pressure of 100000Pa.

COMBUSTION CHAMBER - divided into three sections, the two conical ends and the parallel centre section. Gas temperature in the conical ends was set to 1700K and pressure 150000Pa. Gas temperature in the parallel "heart" of the chamber was set to 1900K and pressure to 200000Pa. I hope this adequately represents a combustion!

TAIL PIPE - divided into four (almost) equal length sections. The gas temperature in the one nearest the combustion chamber was set to 1600K with the others "decaying" 100K per section as they got further from the combustion chamber. All pressures 100000Pa.

WALL TEMPERATURES - for all sections these were set half way between the contained gas temperature and 300K (ambient).

ATMOSPHERE - the pressure to both the left and the right of the engine was set to 100000Pa.

I'll now attach each of my "interesting points" in turn. This I'll do one attachment per posting, because my PC has already crashed on me and I lost everything I was typing. This way I only lose one attachment and what I say about it (I hope).

ATTACHMENT 1 (time = 0.00000)

The initial "bang", with pressure high in the combustion chamber and ambient elsewhere.

[toakreon]

ATTACHMENT 2 (time = 0.00023)

The forward moving pressure wave reaches the front of the inlet.

[toakreon]

ATTACHMENT 3 (time = 0.00050)

A rarefaction wave reflects from the front of the inlet, though it's not a very potent one and "disappears" as soon as it hits the combustion chamber.

[toakreon]

ATTACHMENT 4 (time = 0.00065)

The rearward moving pressure wave reaches the back of the tailpipe. The vestiges of the reflected inlet rarefaction wave can still be seen, just before it disappears into the combustion chamber.

[toakreon]

ATTACHMENT 5 (time = 0.00102)

A rarefaction wave is obviously reflecting from the end of the tailpipe and heading back towards the combustion chamber. As mentioned before the rarefaction wave in the inlet has now almost gone.

[toakreon]

ATTACHMENT 6 (time = 0.00179)

The rarefaction wave in the tailpipe is now well on the way back to the combustion chamber. It seems to be almost "deepening as it goes" - though I can't understand why.

[toakreon]

ATTACHMENT 7 (time = 0.00218)

The tailpipe rarefaction wave reaches the combustion chamber. By now we're getting the beginnings of a genuine inflow along the inlet (see mach graph).

[toakreon]

ATTACHMENT 8 (time = 0.00264)

Some time later there's a BIG inflow to the combustion chamber from the inlet, and also a definite inflow to the combustion chamber from the tailpipe as well (mach graph).

[toakreon]

ATTACHMENT 9 (time = 0.00350)

Combustion chamber pressure is now rising and a BIG inflow from the inlet persists. The air in the inlet and the VERY FRONT of the combustion chamber is reaching quite a high density, but that high density doesn't appear to reach back far into the combustion chamber.

[toakreon]

ATTACHMENT 10 - the last! (time = 0.00508)

The combustion chamber pressure has peaked at about 1.3 bars right along its length, though the gas density is far less even - high right at the front of the combustion chamber but falling to a much lower value for the rest of it.

From this point on the simulation just oscillates down to a steady state - so I guess if anything interesting is going to happen it would have done so by now!

[Rossco]

Hey Mate, Take a good look at some Temperal graphs!
I think that you will pick up a lot more from them.

They are represented by a coloured line per position input on your data sheet. (it has been a long time since i played with Uflow!!!! off the top of my head here)
The respective graphs then, for pressure, velocity, mass flow, etc... are then over the total run time.
Adjust this run time and mesh size to suit your engine, or, to focus in on a particular portion of the cycle.
Dont just rely on 3 or more ocsilations, this is a little broad for my liking. I tend to bring it up in focus, with a small mesh and short run time.
Remember that you can effectively start the simulation in any part of a cycle. Its most interesting to look at the induction "stroke", so start off with a lower pressure in your CC.

You surely are right into it!
(sorry if i missed something in your previous posts... i have to admit that i did not look at all, or any in detail... maybe time later)

Rossco

[toakreon]

Oh ... Silly me (D'oh!!)

Suddenly realized this little lot was next to useless without also posting the dimensions of the "Squirrel"

Here is a very rough-and-ready diagram (drawn by my own, fair hand ... ;) )

John

[toakreon]

Damn - I'd forget my head if it wasn't attached. All dimensions in cm ... except right at the bottom (underneath the linear scale) where they're in meters to match the scale in the UFLOW1D graphs!

John

[larry cottrill]

John -

My take on the curves was that they're not bad, but the velocity node seems a little far forward. Now that we have the dimensions I see why. I think you should either lengthen the intake or shorten the tailpipe.

There is some point near the front of the chamber where the velocity will be near zero throughout the whole cycle - you can look for that by working the vertical slider at the right side of the graphs and watching for this 'nodal point' on the Mach Number graph. When you find that point, try to get your engine to about 1/4 of its length from there to the intake end and 3/4 from there to the tail end. This is inexact, of course, but you should find some layout close to that ratio that runs as well or better than what you have.

Also, once you find the nodal point and have what you think is a good layout (i.e. you think you'll only need minor tweaking from here out), make sure one of your 'Output Points' is located there and use the 'Temporal' plot Rossco mentioned. Try to get a time span that runs about five or six complete cycles, and select 'Other Points' as follows: One just inside the intake edge, one at the nodal point and one just inside the exhaust edge. When you have a good design, the curves will not 'run down' rapidly, but will 'run out' with only minor degradation in each cycle. This is the natural resonant 'ringing' of the air in your engine. It is highly dependent on shapes and lengths, AND on your gas temperature estimates in the different sections of your engine.

Looking at the engine over time in this way probably gives you the best judgment you can make about its overall general character. This view is ideally suited to starting with low pressure (engine ready to breathe), which you enter as values that smoothly deepen from the ends into the chamber (as smooth as you can make it, that is). Running over several cycles lets it smooth out fully and form up the natural curves that your engine layout and temps bring about.

L Cottrill

[toakreon]

Thanks Larry,

I'll have a go with a longer inlet and look for that nodal point ... :)

John