Non-continious fuel injection

[ace_fedde]

Check out this research paper:

Improving the performance of a valveless pulse combustor using unsteady fuel injection Miller.pdf

(520.81 KiB) Downloaded 1039 times

With non-continious fuel injection they reach efficiency increase factors of 2 till 7!!
The idea is pretty logic: only inject fuel when there is a pressure peak, because then combustion is most efficient and
-pulse- productive.

btw, Who can help me find Kentfields research papers (the UAV-PJ I already have)?

Fedde

[metiz]

Good post, Ill have to read up on that soon. I found this "paper" in my folder, might be usefull somehow

[Eric]

While the end results are no doubt better than continuous injection, they make a lot of assumptions in that paper, and display things in a disingenuous way.

Saying the thermal efficiency is increased by a factor of 7, along with a 30% margin of error doesnt tell you anything specific about the engines performance.

I also doubt if they actually graphed the baseline data in any way in the last graph, since the efficiency over the entire throttle range is not a constant for pulsejets. If the baseline run followed the same shape curve as the timed injection, which it should to some degree, it would shift the factors to being much closer to 2 relative to any given fuel flow rate.

According to their definition of relative thermal efficiency, and their graph data, the engine would produce more thrust at low throttle than it would at max throttle. If it were a factor of 3 at low throttle shifting to a factor of 2 at max throttle, it would have at least been more believable, as it would have given an increasing thrust output as you add more fuel.

As far as thrust is concerned, timed injection will never more than halve the TSFC with a perfectly implemented system with an optimized engine, and the data is either displayed inaccurately, or 'tweaked' to make for a much better sounding abstract, and more funding.

Eric

[ace_fedde]

Eric,
Point taken. But as you will see later in this thread, I am more interested in another (expected) aspect of this technology. Any efficiency increase is of course a goal, but the reliability of the numbers is not (yet) important to me. Hopefully, forum experiments will judge in the future.
Btw, do you have any Kentfield papers? (exept for the UAV-PJ)

Metiz,
I would go for a common-rail (diesel) injection system. As you read in the paper the researchers use automotive (gasolin I reckon) injectors and already face problems with the injector's minimum cycle length.
Common rail injectors (specially the piezo ones) are much faster.
Btw, do you have...

Fedde

[Eric]

Somewhere on one of my hard drives I do, look around the forum here, pretty sure mike posted some of them a while back.

Expected aspects of the technology? Not trying to make a PDE are you?

[ace_fedde]

I would be in the wrong forum then, wouldn't I?

Nope, the expected (by me) aspect is "taming the beast" or "PJ management system". Sounds complicated but might be simpeler then expected (by others).

As we have seen in figure 10 the PJ becomes pretty "frequency stable" from the non-continious fuel injection. Also we see in figure 8 that the frequency alters if the fuel is injected earlier or later then the optimum.


Now let's in mind extend Millers experiment:

Imagine that we have that PJ running in little less critical fueling mode (130 degrees pulse width and starting point at 165 degr.) We measure, for example, a resonance frequency of 220 hz.

Now let's block the injectors' (electrical) pulse generator on that 220hz and cut off the pulse detector that normally steers that pulse generator. Now the injection system injects without any correction, or correlation with the PJ, 220 times a second.

Everybody would expect mismatching to occur and the PJ to stop resonating, but ..
Look carefully at figure 8.
If the injection moment starts mismatch and to be (a little) late, the PJ will start to resonate at a lower frequency. That makes the cycles longer, which then again causes the injection moment not to be (so) late anymore in the next cycles. This will take place till the injection moment, due to the lower frequency, starts be to early. Now the frequency will increase, and the opposite will occur.
If that whole proces stays within those soft and hard margins from figure 8 (that's why I took the higher pulse width) the PJ will become stable at 220 hz.

If this is really working out that way it will open great possibilities. Think of out of phase operation for multiple PJ's, valve actuation (here you go SR71!) for valved PJ's. Just couple it to the injectors pulse generator and there you go!

Of course it all has first to be proven to work. (SR71?)

Fedde

[Eric]

I just made a post in the valved section on one of SR's threads, about natural frequency variation and modulation.

Yes you could force the engine to run in a way that it does not like to some extent, the frequency of the engine actually changes along with the magnitude in a harmonic fashion all by its own, but the real question is, if you try to make it do things against its will, how will the performance be affected, and with the fuel timing bouncing up and down bordering on the edges of flaming out, how will overall stability be affected, how much will throttle response be slowed, ie needing a 'spool up time' so it doesn't flame out, and will a breeze cause it to shut down.

[ace_fedde]

Eric,

I read that post in SR's thread, very educative, thanks.
A question, that we probably cannot answer without experimenting, that immediatly rises to me is: Is steady fueling an inaccurate form of fueling, causing more frequency jumps than non-continious fueling (NCF)?
Appearantly steady fueling causes a surplus of available fuel, which surplus might cause the jumping (sometimes the surplus instantly deflagrates, sometimes it is left for the next beat)
The next question is then:Does NCF force the engine to run in a way that it does not like to some extent? Or does it force the engine to run in a way that it most likes, and does steady fueling prevent the PJ from running in that way (running like a clock )?
Then there might not be a bouncing in fuel timing, as you suggest, since the motor runs much more accurate (hopefully).
I think only experiments can give the answers. Unfortunatly my workshop, due to a situation, isn't available to me for a long time already.

Fedde

[ace_fedde]

Btw.,

These are the threads that Eric and I mention:

viewtopic.php?f=2&t=5527
viewtopic.php?f=2&t=5534

Fedde

[John Hasler]

ace_fedde writes:
viewtopic.php?f=2&t=5527
viewtopic.php?f=2&t=5534

Interesting discussions! Was anything ever built? Lots of ideas there, some (but by no means all) of which I had already thought of.