Liquid Phase LPG Injector Calibration 2

[BogRob]

Gentlemen,

Would anyone be able to share a method to measure actual injector pulse width (msec) vs volume flow at a fixed fuel pressure for a liquid phase (lp) LPG injector?

I have come across eight Siemens LP LPG injectors used in the Icom JTG systems which I am planning to use in my own application. The rationale behind not simply using one of their systems is I can see many shortcomings where fuel composition and density are not accounted for, nor am I over the moon about any ‘interceptor’ style engine management.

I hope to learn a lot during this project except for it to begin I either need the data for these injectors or to measure it myself. My current ideas to measure the data are as follows:

1. Using a bottle of N2 to regulate the fuel pressure of the LPG, systematically cycle the injector (with nozzle to atmosphere) at various fuel pressures and supply voltages whilst recording the mass lost from the LPG and N2 containers by means of scales.
The preconceived flaws in this system are the unavailability of a scale that would be able to accurately weigh both containers with resolution to capture the mass loss rate. Mass of fuel loss would also be needed to convert into volume loss rate.

2. Similar process to number one; however in this instance the LPG would be captured into a sealed pressure capable vessel of known volume. Based on the temperature and pressure of the fluid and the known volume, the liquid volume could be calculated.
Again the inherent flaws in the process would be that the N2 pressure regulator would need to compensate for the decrease in pressure differential as the capture vessel pressure increases. The volume of the reference line would need to be accounted for.

3. Use a test fluid that is liquid at room temperature and directly measure the liquid volume whilst cycling the injector as in process one.

I am in favour of process three; however I am unaware of a suitable test fluid which either shares strong physical characteristics of Liquid Phase LPG or that can has known correlation factors to correct the collected data so that it represents injection of Liquid Phase LPG.
I look forward to all future responses and thank you in advance.

Regards,

Robert Flynn
Mech Eng (Hons)

 

[patprimmer]

Method 4

Inject the LPG into a container that is under high enough pressure to keep the fuel liquid. Measure volume increase and do your sums to correct for temperature, SG and pressure difference.

You could even develop your own correction factors by deliberately changing the parameters one at a time and plotting the results.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[Lou Scannon]

Or, if keeping the LPG in liquid phase post injection proves to be difficult, just collect it in a large fixed volume gas tight container that you weigh pre and post test (cfor added accuracy you can corrrect the container volume for pre and post pressure due to container expansion under pressure). Be sure to use a controlled composition test fluid or do a composition analysis so you know what the injected mass consists of.

 

[gruntguru]

Same as method 4 except weigh the container. If the high pressure container weighs too much for the scales you have available, reduce the force using a "beam" type counterweighting system.

You could do this with method 1 too. BTW why bother with the N2? Just warm the LPG bottle to a higher vapour pressure than you need and regulate down to the test pressure.

Engineering is the art of creating things you need, from things you can get.

 

[TStaples]

Robert,

My background includes experience designing and manufacturing gaseous propane injectors (low pressure gaseous at the inlet). The method I previously developed to look at shot to shot repeatability on those, would also work for your measurement needs with liquid propane inlet.

We set the injector up to spray into a fixed and measured volume that was held at a very precise temperature using a water jacket arrangement around the downstream volume. The water jacket was fed by a temp. controlled coolant loop with LOTS of coolant volume. The inlet pressure was set such that the appropriate pressure drop was developed across the injector. A very sensitive low pressure range transducer was used to measure the small downstream pressure increase. As each shot was fired, a resultant rise in downstream pressure would result, and we used the ideal gas law, to convert to a mass flow. It would be easy enough to swap terms to get a resulting volume based on density at a particular pressure /temp. (like STP, for example) We also used a very precise temp. measurement system inside the volume, to measure the gaseous propane temperature, but found that it didn't vary enough from the jacket temperature to worry about, on a relatively small volume. As I recall we were in the 50cc range on internal volume.

In your case, you could do this for longer durations. That would require you to use an electronic pressure regulator to maintain a constant pressure drop during the flow measurement run, based on the pressure rise downstream. Or, as an alternative, you could just let the downstream pressure rise as the injector flows, and back calculate it at each measurement point, to account for the change in pressure drop with time. That would be trickier, but doable. The secret is to get very good control of temperature, so that term doesn't dominate the pressure term, leading to error.



-Tony Staples

http://www.tscombustion.com

 

[TStaples]

After writing the previous post it occurred to me that a simpler method might be the following: Use a liquid flowmeter on the inlet, measure the resulting volumetric flow, covert that to mass flow, and then use the ideal gas law to calculate the resulting volume at the downstream temp and pressure.

For example: Set up the test, run 1000 pulses at the required pulsewidth and inlet pressure, record the displacement of the flowmeter, and then you can calculate the average liquid equivalent volume per pulse. Similar to my earlier pulse, you can then convert that to gaseous volume per pulse at a given set of conditions using the ideal gas law.

I really loved working on compressible flow injectors. The ideal gas law made mass flow conversions between points, and even between test fluids, so easy. For example, we calibrated injectors on Nitrogen, to get specific mass flows on Propane. Calibrating gasoline, or Diesel injectors, is much more complicated because they are typically calibrated on Stoddard Solvent, or white spirits. Density and viscosity creep into the mix.

-Tony Staples

http://www.tscombustion.com

 

[TStaples]

Rob,

I should have gone back and reviewed your OP before writing the second method. The second approach I wrote about assumes you would use Liquid Propane as the test fluid. Your investigating a suitable calibration fluid indicates that you likely don't have access to a hazardous flow area. Is the goal to perform the measurement, or to ultimately develop a production process that is more fundamentally safe?

-Tony Staples

http://www.tscombustion.com

 

[BogRob]

Good morning guys,

I believe method 4 is the most elegant approach for a number of reasons:

1. There is no phase change, the LPG remains as a liquid which allows the N2 to be purged off as the volume is displaced by the LPG. This simply allows for delta mass to be calculated and converted to volume by measuring the fulid temperature pre-injector.

2. The N2 gas is usefull for a number of reasons. Using it to maintain a constant pressure pre-injector means there will be no phase change on the tank side, consequently the flid temperature will have less variation.
The N2 also gives me the ability to measure injector pressure differentials below the vapor pressure at ambient temperature conditions without chilling the fluid.
Additionally, with two regs, I can increase the pressure in the capture vessel to get a positive flow of LPG back into the storage tank and have very little loss of LPG. This is primarily important due to an environmental, health and safety point of veiw.

In preference to the above method would be a test fluid. Finding a fluid which is inherently safe, is liquid at ambient temperatures and shares very similar dynamic viscosity characteristics to LPG is unknown to myself.

Mr Staples:

Your spot on I do not have access to a hazardous flow area, this experiement will be conducted in my garage. Honestly I would prefer if someone was able to just hand me the data required for this injector, however contacting Seimens and the local distributor for Icom has not led anywhere. If anyone has any information regarding these injectors I would be very appreciative.

 

[patprimmer]

Terry has a good point.

LPG is quite a dangerous material to handle.

While method 4 in theory keeps it contained, LPG is actually quite difficult to contain as pressures are high and it's molecule is small, so it will diffuse through materials that are impervious to hydrocarbons that have higher boiling points.

A very small hole or gap can leak a large amount in a small time due to the very high pressures.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

http://eng-tips.com/market.cfm

for site rules

 

[franzh]

The current aftermarket suppliers for LPFI engines utilize flow matched injectors based on a nominal temperature pressure index. Most of the LPFI systems do not use an "interceptor" module, but use the vehicles PCM in adaptive trim mode. There are only a few of these DEKA injectors, most of them use calibrator orifices to fine tune the injector flow rate for the engine and power requirement.

The base injector flow capability is determined by fuel density, pressure, caloric value. One of this forum's members is/was a part of the liquid injector design team and may chip in his vast knowledge.

Franz

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[BogRob]

franzh:

I believe you are right. The only aftermarket product I have been able to find generous amounts of information about is the Icom JTG stuff. Initially I believed it was more sophisticated than I do today, that the Icom PCM varied the LPi injector pulse width based on fuel pressure and temp inputs but you are correct, the only metering method is by the calibrator tip and OEM PCM injector output.

These product deficiencies are part of my motivation to go ahead with this project, I have read a number of papers that empirically determine that there is a considerable variation in lambda from targets as fuel pressure and temperature fluctuate. Additionally, these aftermarket systems do not calculate the tank fuel composition which further exacerbates the measured lambda from targets.

As I've noted earlier, I have very little information about this injector or the other Siemens Deka LPi injectors. I would be in preference of being provided with this information rather than attempting to measure it. Franzh if you are aware of the member on this forum who may be able to assist could you please lead them to this thread, I would be very appreciative.

Thanks guys for all your input.

Rob

 

[jzabs]

Hello BogRob,
This is something that I had to develop when the company I work for was developing its own liquid phase LPG injection system. My solution was to develop a test rig which essentially had 2 fuel circuits. I had a large (90 litre) main LPG tank which housed an automotive LPG pump, this pumped LPG to an external adjustable back pressure regulator, the relief flow of which went back into the LPG tank. This allowed me to provide LPG to another circuit at a pressure above its saturation pressure. This pressure was monitored by a pressure transducer. The second circuit had a much smaller LPG tank, just big enough to house a second LPG fuel pump. The LPG fuel rail was supplied fuel from this small tank and the return fuel from the rail (it was a return fuel system) went back into the small tank through another adjustable back pressure regulator. Fuel rail pressure was monitored using another pressure transducer. The small tank was supplied fuel from the first circuit, through a small coriolis force flow meter. There was also a valved circuit from the small tank to the large tank to purge vapour from the small tank circuit.

The basic operation of the system involved turning on the main tank pump and opening the small tank purge valve, which started to fill the small tank. Once the small tank had sufficient LPG, its pump was turned on to circulate LPG through the fuel rail and purge it of vapour. After a while the vapour would be completely purged from the small tank and its circuit and the purge valve would be closed.

The injectors would then be pulsed and the flow meter (between the 2 circuits) would measure the amount of LPG injected. The flow meter was essentially measuring the amount of LPG needed to replace the fuel flowing out of the LPG injectors. The LPG injectors were mounted into another cylinder used to simulate a engine manifold. The LPG vapour was removed from this manifold and consumed in a burner outside the test cell.

The whole thing was controlled by a reprogrammed version of our LPG computer so that I was using the actual drivers and controls for the injectors that the final product would also use. It controlled pulse width, number of pulses, injection frequency, as well as measuring all the pressures and temperature at various places in the circuits. The controller also shut the system down if pressure and tempertures got outside certain boundary conditions.

Once I got it up and running I found it to be very useful to characterise the liquid phase LPG injectors we are using.

I used a re-jigged version of the system that mounted the LPG injectors in the main tank to run a durabilty test on the injectors using LPG as the flow medium.
Regards, jzabs

 

[turbocohen]

Franz emailed me asking to put in some input.. I will keep it simple.. It looks like you are working on developing a solution for a problem that does not exist.

Before I go further, until you are well versed with the hazards of working with saturated liquid hydrocarbons.. DO NOT SCREW WITH LPG IN YOUR GARAGE! ESPECIALLY LIQUID PHASE!

Using P/T lookup tables it is not too tough to estimate what is being pumped through the fuel injectors. The systems from 15+ yrs ago used temp and pressure sensors and open loop operation was pretty well worked out, well enough for sulev emissons compliance anyway.

Initial injector run-in was done with Stoddard solvent which does not correlate well due to viscosity and other differences. All calibration and off vehicle measurements were made using N-Heptane because it correlates well with LPG orifice flow and dynamic flow. Opening and closing times at the same pressure and temperature ranges also correlates well. I performed a lot of tests that discharged fuel into other containers to verify correlation and the results more than satisfied our quest for using fuels that remain liquid at 1 atm.

One interesting dyno experiment showed that if an LPG calibrated injector was put on a dyno gasoline engine at 500Kpa that the same calibration sans cold start enrichment was not too far off to start and run closed loop although tweaks would of course be needed to dial it in. The Deka LPI injector was originally a Viper v-10 iteration modified for 400psi operation. The later versions were modified to reduce leakage and to reduce hot restart issues, which are by far the greatest challenge for LPI system designers. Deposit tolerance is also a big issue we improved on but hot fuel handling is 95% of the problem when designing lp fuel systems within a reasonable cost. If you want to accomplish something useful convince the LP sellers to put deposit control additives in all motorfuel which will considerably reduce valve leakage and Qd-shift in the field. Then come up with a low cost variable speed fuel pump that can momentarily boost fuel pressures 5 bar above tank pressure at 150 LPM while cold cranking voltage drops to 6 volts at the pump.

Not trying to rain on your parade friend but it appears you are trying to reinvent the wheel.

 

[BogRob]

Gentlemen,

My apologies for the belated reply, I had stopped checking this thread a few months ago once it became inactive.

Jzabs thank you for your input, it is interesting to hear the different methods used to realize a solution to this problem. From the commencement of this project I have acknowledged that there are two fundamental approaches; 1) to measure the mass captured post injector or 2) the mass loss from the supply vessel, obviously you chose the latter. This solution was similar to my original proposal, however I was to weight the supply vessel over time and average the mass flow over the number of injector cycles. The problem here was economically finding a scale with high capacity and accuracy to achieve a reasonable discrimination ratio.

turbocohen: I fundamentally agree with everything you have said, however in my original post and the last response to franzh I outlined my motivations.

Thank you for confirming my suspicions of n-heptane as a suitable calibration fluid, although the dynamic viscosity is triple that of propane it is still so small that I theorised there may be little effect on measured flow rates. If you don't mind me asking, how well correlated were the results? Are you able to comment on the correlation ratio 'n'?

The Ford Australia Motor Company debut their monofuel Lpi Falcon last week at Australian International Motor Show. The fuel system was developed by Orbital which as far as I can tell are utilizing Vialle products. In the below link you can see an engine bay image and make out the green type injectors. They have made some efforts to reduce heat transfer by constructing the fuel rail from what appears to be a fibre reinforced plastic, this method would also be more cost effective in the manufacturing the rail.

http://www.caradvice.com.au/125785/...-at-australian-international-motor-show-2011/

Once again thank you all for your input, what I am doing is nothing new, however the data I require is not available to me. If someone was to supply the data I would be very appreciative.

Regards,

Rob

 

[turbocohen]

Since I may be bordering on disclosing proprietary info, will have to limit some of the how to re calibration. Re the Falcon, yes that is FRP and the green overmolded injectors are revised Deka II LPG injectors.

As for determining fuel compo, using a simple P/T lookup table to compensate has proven to do well enough for sulev certifiable systems. Fiat, Ford, Kia, Daewoo and others have used this strategy for a decade and Chrysler Canada for 2 decades. There was a Chrysler Canada guy who went to work on ford systems with a supplier who deserves a lot of credit for resolving complex problems with simple solutions.. Over 15 yrs ago he had the benefit of having hundreds of fleet vehicles on the road to garner LPI experience from, many of them racking up half a million KM's.

On a side note, while working for my former employer, I met with Korean oem customers and taught them how to design and build their own injection systems for butane/lpg before I left the company with the understanding beforehand that our injector was the only one capable of meeting their unique requirements. They put their own system together within a short time that has now been in regular production for years and may come to the US at some point with revisions for higher than US tytpical vapor pressures. I recieved their lpi assemblies for review and was impressed with the degree of attention to detail and the overall quality of their work. (They listened well and took good notes!) The Korean market uses primarily butane with is very easy to handle re hot restart and their monofuel system can deliver any mixture of gasoline and butane if there were to be a butane fuel shortage. (ANY lpi system if properly designed with enough boos pressure and p/t lookup tables that can be utilized to vary fuel boost pressure can do this and I demonstrated this on my test mule several times.. 50/50 gasoline/propane no problem.) This to my knowledge may be the highest single volume market for lpi in the world.

 

[BogRob]

Turbocohen, regarding the injector calibration I understand your situation, this is the exact motivation for my need to calibrate the injectors.

I fully understand the P/T look up tables, it is a very elegant strategy to determine fuel composition. I have read numerous papers by a few Korean's who have documented their theoretical vs emperical results using P/T look up tables.

It would be some interesting refueling infrastructure for a butane/gasoline mix. Australia's Autogas requirements for vapor pressure are 800/1530kPa @ 40c year round, unlike Europe there is no seasonal change. Below is technical document from Shell outlining their Autogas general specifications:

http://www-static.shell.com/static/aus/downloads/fuels/autogas/tds_shell_autogas.pdf

I have read some research that measured typical Australian retail autogas particulates to range from 1 - 5 micron, in your experience what particulate size can be tollerated by the Deka II injectors? Do you have any first hand experience with olefin formation on injector tips?

 

[turbocohen]

I supplied my own pt tables to a dozen koreans engineers but they had a tech with little in the way of acedemic exp who "got it" first and helped translate.. Hope they promoted the guy in spite of little formal education. Suburban Propane provided me with the contact who set up a test run of different compo's that we used as a benchmark. The variety delivered in the field was represented but we also added data points using pure butane and pure propane along with propane with small amounts of methane to aggravate hot soak profiles.. yes there can be some "thanes" in there that can hamper hot fuel handling.

As far as butane/gasoline that is a cinch, in fact there is a good amount of benefit running some gasoline with additives using oil base to clean injector tip deposits that are typically sulphur, olefin or a combination of the two along with micro particles that adhere to them. When we ran into severe injector leakage issues a long while back I experimented and discovered that >99% of the leaks that occured with injectors that previously passed leak tests within spec had deposits where frustum meets cone i.e., where the valve needle meets the valve seat. We worked with an additive manufacturer who employed a retired chevron employee and an lpi specific amine type additive that was not oil based proved quite effective at preventing deposit buildup, enough to clean up leaking injectors and to keep them clean. Olefin deposits are not too tough to clean up but sulphur is a pain in the as_ to clean out once it embeds on valve seats.. another proprietary feature unique to the Deka LPI design involves metalurgy compatible with lp.. sorry, can't go there in this forum, yes there was a ton of $$ spent before this was resolved, no, can't disclose more cept to say the lpg industry still needs to adopt a motor fuel spec that includes at a minimum enough detergent to keep injector innards clean. It is not a cleanliness issue, it is the normal chemistry, heat, material activity that needs to be accomodated. External deposits are an issue as well. The end of the injector is exposed to hot gasses that siphon back up out of the combustion chamber and they react with some hydrocarbons to form sometimes unmanageable depostis. fortunately the edge orifice is self cleaning.. to a point. Again, deposit control additives are needed and it is up to the lp motorfuel providers to adopt this.

As for contaminants, the Deka LPG has a filter on the lower inlet that is good down to 30 micron. The upstream filter needs to be good for <20 microns preferably BUT it has to be a magnetic filter to capture metals that find their way into the magnetic circuit of the injector. The armature and pole piece are wet with fuel and particles can permanently magnetize enough to wreak havoc with this or any injector design. Why this is more of a problem with lpg has to do with steel transport and storage tanks and other ways magnetizable materials find their way into lpg. I put magnets in my lp tanks and upon inspection sometimes they are blooming full after just a few months of use in the field. Imagine what some of them look like after 350,000km.

 

[BogRob]

Sorry for the delayed reply, been extremely busy with work this past week.

I understand the logic behind the butane/gasoline injection strategy, my previous comments were regarding the filling infrastructure. I have not seen a commercial gasoline bowsers which is able to fill a LPG pressure vessel up to a tank pressure of 3.3mPa. Not that I have ever looked for one either.

Thanks for the info regarding the magnetic particles, I have a suitable 10 micron cellulose filter upstream of the pump but never considered material from the storage vessels structure to contaminate the fuel. This features I have not seen on current consumer products.

I have been in contact with a local lab whom primarily analyse fuel composition for industry, interestingly Shell Australia have begun removing the olefin's from Autogas for commercial use. Sulphur on the other hand I have no information other than Australia has very relaxed regulations including gasoline which is proving to be a problem when Euro 6 emissions are enforced.

I have listed a number of papers below which I've found to be useful references, coincidently they are all written by Korean's. Please let me know if you have any recommendations on further reading.

Effects of injector leakage on liquid propane injection engine performance.
Gyeung Ho Choi, Seong Keun Shin, Seok Choun Bae, Yong Jong Chung, and Sung Bin Han.

A study on the injection characteristics of a liquid-phase liquefied petroleum gas injector for air–fuel ratio control
Hansub Sim, Kangyoon Lee, Namhoon Chung, and Myoungho Sunwoo.

Experimental analysis of a liquid-phase liquefied petroleum gas injector for a heavy-duty engine.
Hansub Sim1, Kangyoon Lee, Namhoon Chung and Myoungho Sunwoo.

Injector control logic for a liquid-phase liquid petroleum gas injection engine.
Sungwoo Cho and Kyoungdoug Min.

 

[teknix101]

Hi Rob, I'm now retired however I built an LPi system and have quite a few bits and pieces left over that I can't bring myself to just throw away, in particular DEKA2 injectors and a mould to produce the injector "cups" with some cute design features. I gather from the rules that direct contact is discouraged, so try looking me up on FB.
Nick de Vries (Cairns)

 

[BogRob]

Hi Nick great to hear from you, I'm definetly interested! Maybe you can share some of your experience in this thread?

Btw does your dog wear a hat and sun glasses?