Synchronous Injection

[John Hasler]

Has anyone here studied synchronous injection? Dudley Smith argues persuasively in his thesis that it should improve performance and reduce fuel consumption substantially but he did no experiments: just CFD. As far as I can tell no one has published anything following up on his work.

Note that by synchonous injection I mean use of something like a diesel injector under electronic control, not just back-pressure modulation.

[nalexandrov]

Maybe a simple valve could be acostically coupled to the pulsejet body? This would be a very rapid moving valve, must use a low mass design.

Otherwise you could use a microcontroller with an audio pickup and have it automatically adjust a solenoid duty cycle to match up with the frequency of the pulsejet. Have a drift setting which goes closed loop by determining the optimal frequnecy based on the amplitude of the pulsejet sound. This way if it gets out of sync the pulsejet will get quieter.

[John Hasler]

nalexandrov writes:
> Maybe a simple valve could be acostically coupled to the pulsejet body? This would be a very
> rapid moving valve, must use a low mass design.

I've seen a design with a little steel flap installed in front of the injector so that reverse flow would force it against the injector opening, but that does not allow any control over timing.

> Otherwise you could use a microcontroller with an audio pickup and have it automatically
> adjust a solenoid duty cycle to match up with the frequency of the pulsejet. Have a drift setting
> which goes closed loop by determining the optimal frequnecy based on the amplitude of the
> pulsejet sound.

My plan is to use a gasoline or diesel engine injector controlled by a phase-lock-loop. Adjusting the pulse width would control fuel consumption. A variable delay would allow the pulse to be precisely timed to occur at the best point during the intake cycle. The PLL filter would stablize operation. Dudley Smith showed that this would substantially reduce fuel consumption but he did only simulations.

The PLL et al. would be implemented in software using a micro, of course. I'm trying to find out if anyone else has done any work on synchronous injection since Dudley Smith published. A Web search turns up nothing.

[Mike Everman]

This approach would certainly work. The hurdle is fuel flow and switching at pulsejet frequencies. I only make (made) smaller motors that were around 200 Hz, which on a vehicle would be equivalent to injectors designed for higher than 12,000 rpm. That's where I stopped with it. It was out of the question for mine unless it was some passive modulation.
James talked a bit about finding appropriate injectors, but I don't think he actually tried it. I'll see if I can get him to chime in.

I'm of the opinion that it needn't have a PLL, just a tunable frequency and the motor will synch up if it's close enough to the sweet spot.

[John Hasler]

Mike Everman writes:

The hurdle is fuel flow and switching at pulsejet frequencies. I only make (made) smaller motors that were around 200 Hz, which on a vehicle would be equivalent to injectors designed for higher than 12,000 rpm.

A casual search indicates that there are electromagnetic injectors used in motorcycles that run at up to 15,000 RPM. Typical automotive injectors appear to be capable of 1ms pulses and 2 or 3ms dead time operating from ordinary 12 drivers. Piezoelectic injectors are faster. Thus 200Hz rates seem achieveable with careful driver and fuel line design. One could also use two injectors firing alternately. New injectors are pricey, though. If I build this I'll have to hit the junkyards.

I'm of the opinion that it needn't have a PLL, just a tunable frequency and the motor will synch up if it's close enough to the sweet spot.

A PLL would allow the motor to determine the frequency of operation while removing the phase noise (of which, judging by how the things sound in the videos, there is a lot). I would, of course, experiment with forcing the frequency: the optimum might differ slightly from the natural.

An advanced version of this system would replace the PLL with a fancy Kalman filter with multiple inputs.

[nalexandrov]

http://arc.uta.edu/research/fluidic.htm - interesting paper on the subject

I really do think that the best way to do this is to aucostically couple a valve or use a piezoelectric actuator near the fuel source.

[John Hasler]

nalexandrov writes:

http://arc.uta.edu/research/fluidic.htm - interesting paper on the subject

Yes, interesting. A valveless valve, somewhat analogous to a valveless pulsejet in operation.

I really do think that the best way to do this is to aucostically couple a valve or use a piezoelectric actuator near the fuel source.

I want to be able to experiment with the precise timing of the injection with respect to the motor cycle. This requires active control. The valve might end up not being located right at the point of injection but automotive-type injectors appear at this point to be the most readily available kind of suitable valve. Some of these injectors are piezoelectric. Of course, a custom built injector would be ideal, but there's no budget for that.

[ace_fedde]

John,

Do you have a PJ to experiment with, which model?
Where are you located? I might have some interesting stuff for the project.

Fedde

[John Hasler]

Fedde writes:

Do you have a PJ to experiment with, which model?

No. I may build one this winter if I decide to go ahead with this project. I need to keep costs down.

Where are you located?

Elmwood, Wisconsin. About fifty miles east of Minneapolis, Minnesota.

I might have some interesting stuff for the project.

Please elaborate.

[ace_fedde]

Do you have a PJ to experiment with, which model?

No. I may build one this winter if I decide to go ahead with this project. I need to keep costs down.

That's often the problem with great ideas and plans. If you're working alone you should be a PJ-expert, owning a number of different working PJ's, before you should start a project like this. Just starting up a proven to function PJ already needs some expertism (through trial and errror), let alone building one. It may sound now like I am a expert like that but I'm not, but a friend of mine is.
I don't write this to disencourage you, but to prevent you of becoming disencouraged

Elmwood, Wisconsin. About fifty miles east of Minneapolis, Minnesota.

I was hoping it was the UK (I'm in the Netherlands). A long time ago I planned a project like yours but I lack the knowledge and experience with electronics and plc's. My lack of experience could be easily fixed through my friend Metiz the PJ builder and designer, also from Holland.
If you were in the UK it would be easier to form a team, to not get disencouraged

I might have some interesting stuff for the project.

Please elaborate.

First of all I have materials like automotive injectors. However I don't have what I would choose for: commom rail diesel injectors, but they are easy to buy on a scrapyard here.

Secondly I have ideas . Mostly ideas that do not forget the fact that PJ's normally operate problematic inaccurate for electronic systems.

At third I have a thread about this subject/project:
viewtopic.php?f=3&t=5535
please study this thread thouroughly, it might already have many answers!

Despite the distance, are you in for a cooperation, with you as electronics/plc/ecu-guy?

Fedde

[John Hasler]

ace_fedde writes:

First of all I have materials like automotive injectors. However I don't have what I would choose for: commom rail diesel injectors, but they are easy to buy on a scrapyard here.

The multipulse ones should be fast enough (especially the piezo types) but I don't like the very high pressure requirement nor the need to recirculate fuel. I see that they are making direct-operating piezo valves for natural gas. Those would be ideal, but are undoubtedly out of my price range. I'll have to see what my contacts in the scrap business can scare up. First step will be to figure out what sort of old junkers are likely to contain suitable parts.

Mostly ideas that do not forget the fact that PJ's normally operate problematic inaccurate for electronic systems.

Not sure what you mean. I realize that there is a lot of phase noise in thses things: you can tell that just listening to them. That's one of the things I hope to improve.

I would probably use an AtMega cpu as I have some (they're only $5) and can be programmed using the Arduino IDE. A pressure sensor should be easy: it's not as if the signal is weak. I think I have enough stuff in my junkbox to build the electronics for under $20. I've also thought about synchonous ignition, but one thing at a time. The first thing is for me to build a motor and get it working. That will probably take me most of the winter. I'm also working on getting up to speed on CFD (Code_Saturne and/or FreeFOAM) so I can do simulations.

Despite the distance, are you in for a cooperation, with you as electronics/plc/ecu-guy?

Sure (I'm a retired electronics engineer) but as I said the first step is for me to build a motor. What would you suggest? I'm thinking it should be the simplest possible axisymmetric design so that it will be easy to analyze and simulate as well as being (relatively) easy and inexpensive to build. It doesn't have to be especially small.

[nalexandrov]

I just ordered a thermojet off ebay. When I priced out my metal stock for making one I realized it was like 70$ in materials not counting cutting and welding. Found a thermojet j3-200 on ebay for 99$.

I have electronics and machining capabilities and am very willing to work on one of these injectors.

[metiz]

What would you suggest?

Sorry for the hijack

It would realy help for the engine to have a low frequency (to help you get your system tuned), self starting and higly tollerant of geometry changes.
I would suggest a medium sized thermojet, maybe linear for easy modifying.

Even if you build the entire engine from 1mm stainless steel (recommended) it shouldn't cost more than $10. I get all my stainless from the junkyard, priced at about 2 euro's per kg, regardless of type. More often than not I find perfect sheets ment for laser cutting, still with a plastic film on one side. There's super shiny SS under there. basically the same quality as new material for maybe 1/10th of the price.

[John Hasler]

metiz writes:

Sorry for the hijack

Not a hijack: the question was general. Thank you for responding.

I would suggest a medium sized thermojet, maybe linear for easy modifying.

How large is medium? I'd like to keep it under 2m or so for ease of transport. I also have some scrap around that might be usable: some 5cm pipe and a couple of 10cm x 20cm stainless-steel pressure vessels. I don't expect my first motor to be suitable for my experiments (though I might get lucky). It's primarily for learning. I'd also like to avoid fabricating from sheet as I have no tools for rolling.

[John Hasler]

Just found this:
http://cambridge.academia.edu/RobertMil ... _injection
Membership required. Haven't read the paper yet: just the abstract which is included below.


American Institute of Aeronautics and Astronautics
Improving the Performance of a Valveless Pulse Combustor using Unsteady Fuel Injection

Tom Offord, Robert J Miller, James R Dawson, Jonathan J H Heffer
Whittle Laboratory, Department of Engineering, University of Cambridge, United Kingdom

Sam Mason, and Mark Taylor
Rolls Royce plc, Derby, United Kingdom

This paper describes an experimental investigation into the effect of unsteady fuel injection on
the performance of a valveless pulse combustor. Two fuel systems were used.The first delivered
a steady flow of ethylene through choked nozzles, and the second delivered ethylene in discrete pulses using high-frequency fuel injectors. Both fuel systems injected directly into the
combustion chamber. The high-frequency fuel injectors were phase locked to the unsteady pressure measured on the inlet pipe. The phase and opening pulse width of the injectors
and the time-averaged fuel mass flow rate through the injectors were independently varied.
For a given fuel mass flow rate, it is shown that the maximum pressure amplitude occurs when
fuel is injected during flow reversal in the inlet pipe, i.e.flow direction is out of the combustor.
The optimal fuel injection pulse width is shown to be approximately 2/9th of the cycle. It should, however, be noted that this is the shortest time in which the injectors can reliably be fully
opened and closed. It is shown that by using unsteady fuel injection the mass flow rate of fuel
needed to achieve a given amplitude of unsteady pressure can be reduced by up to 65% when compared with the steady fuel injection case. At low fuel mass flow rates unsteady fuel
injection is shown to raise the efficiency of the combustor by a factor of 7 decreasing to a factor
of 2 at high fuel mass flowrates.