Fuel Control

Moderator: Mike Everman

Post Reply
cbromano
Posts: 35
Joined: Fri Dec 19, 2003 6:20 am
Antipspambot question: 0
Location: Salt Lake, UT, USA
Contact:

Fuel Control

Post by cbromano » Wed Jan 21, 2004 4:28 am

I have been fishing around in junkyards for fuel injector parts to fuel a pulsejet with 25lbs of thrust. I have realized that the entire time I have just looking for the pump and the injector. Have I overlooked a way to monitor fuel flow? The magnetic switch in the fuel injector can start and stop flow but not meter it. Any Suggestions?
Thanks-
CB Romano
I don't wish to know everything.
I just wish not to be ignorant.

resosys
Posts: 292
Joined: Wed Dec 10, 2003 11:26 pm
Antipspambot question: 0
Location: Sacramento, CA, USA

Re: Fuel Control

Post by resosys » Wed Jan 21, 2004 7:24 am

cbromano wrote:I have been fishing around in junkyards for fuel injector parts to fuel a pulsejet with 25lbs of thrust. I have realized that the entire time I have just looking for the pump and the injector. Have I overlooked a way to monitor fuel flow? The magnetic switch in the fuel injector can start and stop flow but not meter it. Any Suggestions?
You can modulate the fuel flow with a PWM controller switching power into the coil in the injector. Rescyou posted some links to DIY fuel injection sites. There are some Texas Instrument ICs that are well suited to switching the power into the injectors as well. I'll try to get the PDFs on my site in a day or so. I'm still learning all the PWM stuff, but will publish my circuits once they are built, tested, and working.

Also, what kind of heat will the injectors be subjected to? If they are anywhere near the "glow", they may not work. I haven't found any decently priced direct cyclinder injectors to try on my valveless pulsejets. On a valved engine, it may be easier to keep them cool, depending on where the injection takes place.

Chris

Rescyou
Posts: 100
Joined: Mon Oct 06, 2003 3:45 am
Antipspambot question: 0
Location: Alberta Canada
Contact:

PWM Metering

Post by Rescyou » Wed Jan 21, 2004 7:39 am

If you setup up a closed system with the fuel output of the injector being sprayed into a measuring container, something that could either measure weight or volume and setup a system were you could give the injector a certain number of pulses, you could measure the flow. The more flow time and the more accuracy. You could also reverse the process and have a measured fuel tank where you subtract the difference after the run.

This would only work if you stayed at a consistant cycle frequency though. There are 4 variables, duty cycle (duty per cycle) which gives you two variables, ON-TIME and OFF-TIME the third variable is the total number of pulses or CYCLES. There is also another variable FREQUENCY, thats how quick the pulses are in a single cycle.

Now duty cycles are based on your fuel/O2 ratios. The longer the on time the richer the mix. What I've had to do is figure out the LEL and UEL of the fuel with respect to the performance of the injector and conditions presented.

LEL or LFL=Lower Explosive/Flammable Limit
UEL or UFL=Upper Explosive/Flammable Limit

Once you know the LEL you measure the the output of fuel for say 2 minutes and log it then increase duty/frequency by say 10% and repeat until you reach the UEL. You could draw that out on a graph or extract a basic formula and you will have a basic tool for figuring out fuel flow which then can be programmed into a realtime display.


I created a program to start with a 0 duty cycle and slowly ramp the duty till the nozzle output sustained flame and then increased it from there until it was to rich to burn.

I then created a program to slowly increase the duty cycle starting from the LEL and ending at the UEL with the output of the injector going to a beaker that was put on a weigh scale which was plumbed into a telemetry controller which recorded the data and graphed it out simultaneously. The data then showed me what the the optimum fuel/O2 ratio was.

I then did the same tests and added the forth variable, Frequency into system which much complicated things.. I generated about 50 different ramp sequences, such as increase duty at a rate of 2 and increase frequency at a rate of three... etc. I put the system on auto and let it run for about 4 hours while it performed various ramps while recording the data. In the end I have some workable formulas which showed the best results given the various ramps. I picked the best set of data and ran with it for about 10 mins using a little more precise volume measuring device and divided the total volume by the total run time to get MM^3/Minute.

It sounds horribly complicated but it's actually a pretty easy setup and the programming was super simple with the exception of frequency which has a lot to do with processor speed in some cases. Frequency is also limited by the mechanical aspects of the fuel injector itself.

Then you can add more variables such as intercooling and fuel pressure which may or may not have any effect. This method works on some and doesn't work on other, it depends on what you have.

A decent place to look for pumps and sensors is in the fuel tank itself.


Some decent reading material:

http://www.oreillynet.com/pub/a/network ... 2/pwm.html

http://www.asashop.org/autoinc/dec99/mech.htm

http://www.autospeed.com/cms/article.html?&A=0008

http://www.autospeed.com/cms/A_0349/printArticle.html

http://www.performancetrends.com/Fuel%20Injector.htm

http://www.mech.kyushu-u.ac.jp/lab/ki05/research-e.htm

http://in3.dem.ist.utl.pt/lxlaser2002/papers.asp

http://www.asashop.org/autoinc/dec2000/mech.htm

S.
The mind of a man is the man himself.

cbromano
Posts: 35
Joined: Fri Dec 19, 2003 6:20 am
Antipspambot question: 0
Location: Salt Lake, UT, USA
Contact:

Post by cbromano » Wed Jan 21, 2004 9:09 am

Hmmm- I have very little experience in circuitry like this. It will take me quite a while to fully comprehend what you are speaking of. I was hoping for a simple solution like a manual valve but as the saying goes "anything worth doing is worth doing right." I am reading the links you posted and am trying to make sense of them. Basically I don’t understand how you found the ideal fuel/O2 mixture. I understood the concepts behind LEL and UEL but not how you figured them. As for programming, I understand and a write in C++ but I don’t think that’s what you were talking about. Don't write me off as dumb. Just VERY ignorant.

Thanks
CB Romano
I don't wish to know everything.
I just wish not to be ignorant.

Mike Everman
Posts: 4934
Joined: Fri Oct 31, 2003 7:25 am
Antipspambot question: 0
Location: santa barbara, CA
Contact:

Post by Mike Everman » Wed Jan 21, 2004 10:24 am

Chris and S:
thanks, your fueling dialog has been very valuable! Permanent archive stuff. I can't wait to play with liquid fuel.
Mike
__________________________
Follow my technical science blog at: http://mikeeverman.com/
Get alerts for the above on twitter at: http://twitter.com/mikeeverman

Rescyou
Posts: 100
Joined: Mon Oct 06, 2003 3:45 am
Antipspambot question: 0
Location: Alberta Canada
Contact:

Fuel-I's

Post by Rescyou » Wed Jan 21, 2004 5:30 pm

Here is a decent link to a Fuel Injector info site:

http://www.injector.com/faq.php

http://www.nasz.co.uk/injectorman/HowTheyWork.html

If you know C++ then you're one step up on me as I do everything in various Pic Basic compilers.

So to find the LEL of something you start off with a long off time and a short on time. and work your way up in duty.

On 1 Off 254
On 2 Off 253
On 3 Off 252
On 4 Off 251...

As the duty increases the fuel/O2 mix gets richer and at the LEL the material will support combustion. As you increase the duty again..

On 200 Off 55
On 201 Off 54
On 202 Off 53...

It will get to a point when the mixture is too rich and won't support combustion, thats your UEL point.


I believe the O2 Sensor with a little help from the Mass Air flow sensors in vehicles sense the correct mixtures etc. and adjust the duty as needed.


Generic PWM Algorithm :


START

Get input from the user on the Duty, we will use Duty=5 for this example

Cycles=255

Turn On Injector

Increase Cycle by 1 starting at ZERO and ending at Cycles

If Cycle=Duty then Shut Off Injector

Loop back to "Increase Cycle and increase by 1 until Cycle=Cycles

Go back to the START and do the whole thing again, over and over

So if we look at the data coming out:


#1: Cycles=255 Duty=5 Cycle=0
#2: Cycles=255 Duty=5 Cycle=1
#3: Cycles=255 Duty=5 Cycle=2
#4: Cycles=255 Duty=5 Cycle=3
#5: Cycles=255 Duty=5 Cycle=4
#6: Cycle now = 5 and duty = 5 so shut off the injector
#7: Continue to count upword until cycle=cycles with the injector off

So with a duty of 5 you are actually keeping the injector on for 5/255's of the time.

QBASIC Source file that should be able to run from Windows (Go into RUN and type in "QBASIC" and type in the code:

-----------------------------

Duty=5
Cycles=20
cls

InjectorStatus$="ON"

For cycle = 0 to cycles

if duty=cycle then InjectorStatus$="Off"

print cycle,duty,InjectorStatus$

next

-----------------------------


Generic PWM Algorithm for a ramped up output:


Get Duty for LEL and Duty for UEL

Cycles=255

START

Increase duty by 1 starting at LEL Duty and ending at Duty UEL

Turn On Injector

Increase Cycle by 1 starting at ZERO and ending at Cycles

If Cycle=Duty then Shut Off Injector

Loop back to "Increase Cycle and increase by 1 until Cycle=Cycles

Loop back to "Increase Duty and increase by 1 until Duty= UEL Duty

Go back to Start and do it all again

Qbasic code:

-----------------------------

LEL=5
UEL=10
Cycles=20
cls

For Duty=LEL to UEL

InjectorStatus$="ON"

For cycle = 0 to cycles

If Duty=Cycle then InjectorStatus$="Off"

Print cycle,duty,InjectorStatus$

next cycle

-----------------------------


The Qbasic code should work but I didn't have to check it out.

S.
The mind of a man is the man himself.

Post Reply