Pulse-Jet Operating Frequency Variations

[SR71Fan]

Does anyone out there have any actual measured data on instantaneous
variations in pulse period under normal conditions? I know that even
under ideas inlet conditions, there will statistical variations.
But I would like to find quantitative data.

SR71Fan

[Mike Everman]

I do not have hard data, but you could generate that by using one of the freeware programs like Visual Analyser and use it on some pulsejet videos (where the camera does not move around).
Frequency does not change much with power level, in my experience. At speed is a different matter. Maybe there's a paper on that somewhere. I'll look through my files.

[Rocket Man]

I am not exactly sure what you are asking for. Here is some good data from a Dyna Jet engine running on gasoline. This is research data that a college student did for his graduate thesis paper.

Combustion chamber pressure was measured to be 3.6 lbs.

Thrust was measured at 4.36 lbs.

Pulse rate was measured to be 220 Hz.

If the tail pipe is made 25.5" long frequency changes to 208 Hz.

You get the idea.

[SR71Fan]

Rocketman:

In your example, the pulse rate was measured as 220 Hz.
This is an average of 4.545 milliseconds between pulses.

What I am trying to find out is exact time between successive
pulses on a pulse by pulse basis. How much do any of the pulses
change from the 4.545 millisecond average?

SR71

[Rocket Man]

The research paper said they used a pressure transducer to check the pulse rate. It also said something about variations in pulse rate changed very little during the experements. This is a large pdf file I can not copy and paste here. I can email it to you if you send me your email address.

There is a German V1 in a museum in California. I know a guy that works there they fire the engine up all the time for the visitors. He says the pulse rate is 40 to 44 Hz. The engine starts and idles at about 250 lbs of thrust then they throttle it up to 780 lbs of thrust. They run the engine for only 60 seconds each time they fire it up.

There seems to be a lot of conflicting information online about the V1 engine. I have read pulse rate is anything from 39 to 45 Hz. But who knows it may vary from engine to engine.

[Eric]

It varies by fuel, atmospheric temp, and air speed. In controlled experiments you wont see much variation, but otherwise you will see a +/- 5% shift in frequency depending on conditions.

I never got around to releasing my pulsejet calc 2.0 which graphs the frequency range under a number of conditions, but the 1.4 version you can download from the main site pulsejet section has a frequency spread calculator that will give you an idea of an engines frequency range with gasoline as fuel.

[SR71Fan]

Eric:

The important parameter is the instantaneous (pulse-by-pulse) variation
in the pulse period (1/frequency). If it is actually less that 5% in the
short term, then that is acceptable.

For the reason I am asking the question, refer to "Rotary Valve Pulse Jet Engine"
in the "Valved Pulse Jet" forum.

SR71

[Eric]

Statistically there wouldn't be that much variation over one second with normal running at a set throttle, but most engines have a pulsation magnitude frequency where the thrust may go from a relative minimum to relative maximum 10 times a second. Say like a 4.75 lb engine going from 4.5 lbs to 5 lbs at a frequency determined by a fraction of the operational frequency. The magnitude of the modulation varies by design.

The change in frequency from one pulse to another could be even more than +/- 5% if you quickly increased throttle right as the engine was at one of the relative minimas at a low throttle. If it skipped a cycle and kept going, building up a substantially bigger charge the next cycle, the instantaneous change could be quite dramatic.

Also because of this variation in intake / output there would be an effect on the engines operational frequency, with all the effects interacting with each other constantly bouncing up and down to find a balance.

Problem is once you start trying to force the engine to do things against what it wants to do naturally, you may get a lot of sudden shut downs. Some engines can skip a cycle or two and keep going, although usually if the engine skips a cycle, it will lead to rough running which can cause other misses, which become more frequent build until it shuts down completely. This would become even more pronounced with the rotary valve being unable to instantly adjust like the petal valves.

Overall you could probably get things close enough to run, but it may be one of those things that are extremely hard to fine tune or get to run as well as a normal valved engine.

[SR71Fan]

Hello Eric:

This is regarding your reply to Pulse-Jet Operating
Frequency Variations that is dated Sun Oct 25, 2009 12:03 am

Thanks for taking the time to reply on this topic.

However, it seem that I am miscommunicating or something, because
everyone seems to be answering questions that I am not asking.

My fundamental question regarding operating frequency variations
of a pulse-jet needs to **first** be answered for conditions of
steady-state operation: constant air temperature, constant air
inlet conditions, constant fuel feed conditions, constant throttle
conditions.

It is not my goal to try to force the engine to do things against
what it wants to do naturally.

I would like some insight on what you described as engine thrust
variations occurring at a **fraction** of the normal operating
frequency. In engineering terms, this is often called a
"sub-harmonic oscillation", and this **is** something I am
interested in, especially if it occurs under the steady-state
conditions I've outlined above.

Does anyone have empirical evidence for what can cause such a
sub-harmonic oscillation, or are the causes merely conjecture?
Does the phenomena appear more often with larger or smaller
engines?

Any insight you have would be appreciated.

SR71

[Eric]

Yes it is harmonically driven, and I suspect it happens in all pulsejets, its just a matter of wether or not its on a scale that can be measured.

Im not sure if your question is something that can be answered, even if you had perfectly controlled steady conditions, running the engine would destroy the conditions, any results would seem highly randomized and chaotic.

Maybe when computers get to the point where they can simulate all the interactions of every single atom in a test environment... For right now there are just way too many variables when you have an oscillating system that causes everything else around it to oscillate and heat in different ways which then in turn effect the oscillations of the thing that is causing them to oscillate, and the things around them.

Currently 'reality' is the only thing capable of calculating all those variables in real time with any degree of realism.

[SR71Fan]

Hello Eric:

Here is how I approached the issue of short-term frequency variations

Parameters affecting pulse-jet operating frequency
---------------------------------------------------
Assume: Fpls = a* / (4 * L)

a* = speed of sound inside the pulse-jet (approx. 2100 ft/sec)
L = total engine length in feet

NOTE: "short-term is defined as less than 1 second.
"long-term" is defined as greater that 1 second.

Effects on short-term frequency stability are listed in order
of importance, lowest impact first.


1. The engine temperature stabilizes slowly after engine start-up, so
variations in "L" can be ignored. Effect is long-term only.

2. Air blowing on the exposed sides of the engine tube will cool the
tube. Effect is long-term only.

3. Higher flight speed will change the reverse exhaust influx
pattern (due to the higher rearward vacuum) and will alter the
pulse frequency. Effect is long-term only.

4. Higher flight speed will increase the inlet dynamic air pressure
(ram effect) and will alter the pulse frequency.
Effect is long-term only.

5. Variations in fuel/air mixing during the air induction cycle can
result is significant variations in mean exhaust gas temperature.
Directly affects "a*". Precise fuel injection control can mitigate
this effect, but may not eliminate it as a variable.

6. Wind gusting at the air inlet will skew the air induction cycle,
resulting variations in pulse period. Effect is short-term.

7. Random variations in the reverse exhaust influx will induce
sporatic delays in ignition of the follow-on air/fuel mix.
Possible mitigation for this effect is an "intelligent" fuel
injection timing profile. Effect is short-term.

This form of analysis is referred to as a "qualitative analysis".
The goal is **not** to produce a numeric solution; it is to examine
factors to determine the major contributors to errors. Some of
the potential error sources do not predominate, so the focus
should be reserved for those error sources that will predominate.

Items #1 to #4 can be ignored, as their effects can be tracked
within the control mechanism.

Item #5 can be **minimized** in the design of the fuel control system.

Items #6 could be somewhat reduced by the design of the air inlet,
but a thorough analysis would be difficult

Item #7 is the major unknown, since the ignition of the new air/fuel
mix by the reflected exhaust flame is literally beyond our ability
to simulate, as you have already pointed out.

This method is not as rigorous as your modelling methods, but gives
an indication on which error sources will give the headaches.

SR71

[tufty]

It's not rigorous at all. It's guesswork.

[SR71Fan]

Tufty:

It's not merely "guesswork"... it's **educated** guesswork.

There are many things we can infer about the operation of a pulse-jet that
are based on fundamental physics. If we can set those issues to rest, we can
then focus on what we don't know.

I think Eric will agree that the problem with "modelling" is that we can't be sure
that our model is correct, and, if not. "garbage in --- garbage out".

"Rigorous" in context of the discussion will depend on whether our techniques
incorporate "quantitative analysis" (Eric's specialty), or "qualitative analysis",
which is the initial **engineering** approach if we know that mathematic
modelling cannot be trusted because we don't know enough about all the
interactions in the system. In Eric's world, "rigorous" means that we can
trust our model. My perspective is that "rigorous" in the case of a qualitative
analysis is the confidence that you actually "know" all the things that you
**think** you know, and are willing to admit that the rest is a mystery to
be investigated by experimentation.

So the only way I can confirm that my rotary valve pulse-jet concept will work
will be to build one. And my "qualitative analysis" has shown me what I can
ignore, and what I have to focus attention on.

SR71

[tufty]

Educated it may be, but it's still guesswork.

For example, your numbers 5 & 7, which make an assumption that ignition is driven solely by the drawback of hot gases into the combustion chamber. I feel this is almost certainly not the case. I think you'll find that, for a valved pulsejet, ignition "starts" at the moment the valves open and a new charge of fuel / air mixture enters the chamber, hitting the still-burning mixture that's already within the chamber, especially the boundary layer (see the SNECMA / Messerschmidt work), not to mention the cylinder walls themselves. Remember, this is not a piston engine, where combustion has completed and the cylinder is cleared and cooled before a new charge enters.

I can see problems or misguessing on all the others as well.

What you're talking about is trying to control combustion in a pulsejet, to drive it to work your way. Which is a fair goal, but there are a vast number of other unknowns waiting to trip you up. Flame speed, turbulence, mixing times, injector type pressure and pattern, etc.

Let's assume you have a combustor which has a precisely controlled inlet and fuel injector. I very much doubt that you can control ignition timing, even with an external ignition source, unless you're making the engine fire on one cycle every 2 or so (this might be a reasonable goal for "idling", of course, but effectively involves "shotgunning" the engine and largely ignoring the acoustics) by not injecting fuel every cycle - this is about the only way I can see of scavenging all the hot gas from the CC and cooling the CC walls enough to make your ignition source the "only one".

As Eric says, the only way to model this accurately is by using reality. You might be able to make a computer model and get an idea of what variables are dominant, but in the end it comes down to cutting a lot of steel (and, in your case, code). Guesswork isn't going to get you very far if you're looking for a precision result.

All MHO, of course.

Simon