Propane, lean burn, with no throttle? 1

[obanion]

Came across a couple interesting experiments. One particular thing caught my interest:

________________________
Full experiment here:

http://www.aist.go.jp/MEL/soshiki/ene/nensyou/shakal/CVC/CVCpage1.html

________________________
Also mxiture tests with LPG:

http://www.aist.go.jp/MEL/soshiki/ene/nensyou/shakal/InletFlow/InletFlow.html

________________________

The engine I have has a 8.5:1 CR. Unthrottled (1atm in the manifold), I should have a in cylinder pressure of about 2Mpa, and a temp of about 414C ATDC. According to the experiment, it should still be ignitable at over a 2 lamdba AFR, probably more like 2.5 or more if the gain in lean ignitability up to 200C (as high as the experiment goes) is any indication.

BTW, the fuel system is being converted to a constant proportional injection of propane (computer controlled proportional control valve), just like in the above experiments. After reading the inlet mixing tests, I'll probably use the exact method of injection ("j" jet) that they found to work so well.

Previously, when the motor was a normal throttled gasoline motor, in order to maintain a idle speed of about 1000RPM, the throttle was adjusted get a manifold pressure of ~60kpa (60% of 1atm), with a stoich mixture of gasoline.

So am I right to think that if I took the throttle out, and set it to a mix of about 1.66, it should maintain at or near my previous idle of 1000RPM? Or the removal of the pumping losses of the throttle mean I'll need to run it even leaner, to keep the power low enough to maintain a idle speed? If the required lean mixture is still ignitable, no problem right?

So it would kind of be like a diesel in so ways (no throttle, engine power/speed controlled by fuel), but different (still uses timed spark ignition).

Before anyone mentions it, it's never been tried with gasoline because this approach REQUIRES a gaseous fuel, a liquid fuel has far lower lean burn limits. That's mentioned here, in another great article:

http://franzh.home.texas.net/lean.html

Any big holes in this theory I'm missing?

 

[obanion]

One additional thought. I'd think this would be a easy theory to test. Set up a propane cylinder, a metering valve, and set a static delivery of propane, just enough to get about a 2 lambda mix at 1000RPM, 1atm for the engine. Hook the output of the propane into the intake, turn on the propane, go WOT, and shut off the gasoline injectors. If I'm right it should maintain a set rpm, hopefully close to the target, if not, reduce propane flow until you get the desired rpm. If I'm wrong it will cough and misfire badly.

 

[franzh]

Your experiment sounds valid.

Propane has a lean and rich flammability limit of 2.1%-9.6% which makes throttle-less operation difficult. I have operated engines on propane at L=1.9 but at 30 psig turbo boost. It was VERY difficult to maintain effective combustion at that level and even the slightest stray AF variation resulted in engine failure. Operating an engine at idle or low engine speeds at AF ratios much over Stoich would be very difficult.
Thanks for the page plug.

Franz

 

[obanion]

franzh....what do you mean engine failure? As in the engines FAILED, ie broke parts, or just misfired and/or stalled?

I'm also considering using hydrogen to increase lean burn limits, and combustion efficiency. Perhaps a 10-20% hydrogen by volume mix.

 

[franzh]

Simple rules of thumb with any engine project: Low compression and engine loads allow for wide variations in ignition timing and air/fuel mixtures. The higher the compression (either static or dynamic with boosted pressures) the less tolerant the engine is to these variations. 5 degrees advance may not significantly alter the performance on an engine with 7:1 compression, but at 9:0, there is a significant problem, and at 12:1, you can expect pistons, plugs, and other problems to develop.

One engine project I was involved with comes to mind:
5.6 inline 6 industrial engine on natural gas, 10:1 comp ratio, at 3000 rpm, at 30 psi boost, throttle wide open (throttle-less at this point) was pulling almost 300 bhp at around L=1.65. An Ethane spike in the pipeline gas supply caused a microsecond of fuel to air ratio anomaly which went rich to L=1.16 and the engine seized. Only by examining the rotograph showing the fuel GC pinpointed the cause of failure. Please note that this engine was not feedback controlled and was running in open loop. The Lambda sensors were used for monitoring only. The rate at which the fuel spike and failure occurred leads us to feel that closed loop controls would not have reacted quickly enough. This engine was stressed to the limits (all is fair when doing research!) and the failure was not entirely unexpected, just that we didn’t expect it to happen this way. When the engine cooled, we performed forensic analysis. With the exception of the crankshaft, camshaft, lifters and pushrods, the engine was a total loss.

Using H2 to supplement the fuel is fine, but storage and mixing becomes a problem. Relatively heavy LPG vapor does not like to mix with the extremely light H2. Stratification occurs almost immediately.

Franz

 

[obanion]

OK you were playing around with boost and lean burn.

At L=1.65, the knock threshold was low enough to safely have that mixture and 30psi with a 10:1 comp. At L=1.16 you went right into pre-ignition.

My intention is a dual mode.

Under no boost (cruise, idle) run lean burn mixtures to govern load control.

Once boost begins, immediately transition to a L=0.85.

I'd love to run a HCCI engine, the technology is so promising, but I'm not ready to be that crazy yet. Perhaps a dual fuel, duel operation. Compression ignition with a low octane fuel at low load, switching to a spark ignition with a high octane fuel at full load.

 

[franzh]

obanion:

Dual fueling an engine is not a new concept. Doing it efficiently (and effectively) is your challenge. You must optimize the engine for a fuel (since you cannot change the engine configuration when switching fuels, you must compromise between) which means that the engine will not be operating in optimum conditions for either (or both).

Running rich at boost is where you will encounter detonation. That’s the concept of lean burn (besides the engine efficiency and emissions reduction). Running rich at boost at the ratio you describe will dramatically increase the combustion and exhaust temperatures (remember, we are talking about vapor fuels here, not liquid fuels!)

My recommendation is identifying a fuel then optimize the engine for that fuel. We are currently playing around with finely tuned engines with a certain grade of fuel, then toss in a slight fuel anomaly and watch it go crazy. Just imagine what would happen with an engine tuned for dual fuels when one of them is a bit off spec. The safety redundant controls would be quite complex.

Franz

 

[obanion]

If you are in lean burn mode, you are automatically at lower power. A 1.6L miture won't make the power of a .85L mixture, given the same mass air flow, no matter what you do.

The engine was normally made for gasoline with a turbo (8.5:1 compression). Propane has a higher octane than pump gasoline. If it could run 17psi without knock on 92 pump gas, it should at least do 20psi without knock on propane. There will be, additionally, methanol added in during boost situations. An abundance of octane you could say.

The only time lean burn will be in use is when there is no boost. It's pretty much impossible to detonate 104 octane fuel at 8.5:1 compression with no turbo boost, at any mixture. I can wander the mixture from super lean (2L) to stoich (1L), and cross the "danger zone" of ~1.1-1.2L without concern. But once under boost, if the mixture goes into that zone, look out.

Also consider the flame speed, and temperatures, will be lower with propane than gasoline.

 

[obanion]

I think I didn't make clear, I'm NOT doing compression ignition. I understand the complex balancing act of getting it just right has been achieved only in the lab so far, and then under very carefully controlled conditions. I'm not about to take that on.

I'm sticking with the propane (for sure), full rich at high load, and hopefully throttleless/lean burn for light-medium load.

 

[franzh]

Obanion:

Not sure if I mentioned CI engines, the ones I worked with were all SI. I found that there is insufficient fuel density at high Lambda ratios to run throttleless. When running at high Lambda ratios, where you desire, and an open throttle, at low engine rpm’s (as in a CI engine, with SI controls) I don’t think you will have enough fuel to operate the engine. Only by boosting the pressure will you have sufficient fuel density at elevated air pressure to ignite, and then, only at higher RPM.

I have never had a successful engine operate at low rpm’s at Lambda ratios much over 1.2.

Now, all this said and done, it might make a unique dyno lab project. I would use a motored dyno and a 4 or 6 cylinder inline engine. Use an injected fuel system where you can control the pulse-width on the fly and generate fuel tables based on delta P. A wide-band O2 sensor plus a thermocouple in each exhaust port would help show when you are making power. Once you set the delta P, play around with the ignition timing.

Let me know what you find.

Oh, by the way, 20 psig boost at 8.5:1 at 3000 rpm with 25 degrees BTDC on 104 Propane will indeed detonate, I have several pistons to show for it.

Lastly, I really don’t mean to rain on your ideas, they do have merit. In college, many engineering courses go in great detail about the combustion principals of an Otto cycle engine with a liquid fuel, but either skip entirely or briefly cover the use of vapor fuels. Many combustion principals are changed when using vapor fuels.

Franz

 

[obanion]

You are raining on my parade, but I thank for for it. Save me the pain of learning it first hand.

Tell me what you think of this experiment:

http://carambola.usc.edu/Research/TPCE/TPCE.html

Looks like they were able to get a HUGE amount of range of combustion. I could certainly do what they did, have hot air intake (in my case, pull in air from around exhaust header), then divert to cold air at high load.

But if it really will be impossible to maintain idle speed without a throttle, then the whole idea is out the door.

And how exactly is vapor fuel different than liquid in combustion characteristics? Give me a couple prime examples. I thought gasoline was a mostly a vapor fuel anyway by the time of ignition. I already understand the effects of octane, latent heat, autoignition, and what not. What am I missing?

Also, I don't intend to run propane alone during boost. Methanol will contribute a significant portion of the total fuel under boost. The propane will be by itself only at low-medium loads, once boost starts going, so does some methanol and a little water mixed in.

 

[franzh]

I had forgotten about that study, it was filed in 1996, but performed a couple of years earlier. I once worked with one of the reference scientists. This study used natural gas which has a wider air fuel ratio and works better in far lean mode than propane.

A liquid fuel in combustion requires a finite amount of time to fully atomize, in some cases it never fully does. Injectors that "fire hose" can have low speed misfire due to improperly atomized fuel. This partially vaporized fuel ignites in clumps, instead of a progressively smooth rolling combustion as a fully vaporized fuel or a vapor fuel. Vapor fuel engines typically have smoother very low speed engine performance due to the fuel remaining in suspension with the air instead of dropout as with a liquid based fuel.

One other item I take issue with is heating the air mixture. This decreases air density and in some cases, brings the air and fuel mixture closely to ignition temperatures.

In street operation with parasitic loading and demanding almost transparent operation, I really dont feel that this would work as well as you or I need.

Franz

 

[obanion]

Exactly. Those are some of the prime reasons I'd like to run the car on propane during most driving conditions. I can't flood the motor, wash down the cylinders, it will idle smoother, start easier. But you also sounded ominous like it was intrinsically more dangerous than gasoline at high load, which confuses me. I know propane has lower peak temperatures, slower flame speed, and much higher autoignition temperature. All these things should make it safer on propane than pump gas at any particualr load/timing/speed situation. Not impossible to knock, never meant that, just safer if anything vs gasoline.

I don't like and can't accept the storage problems of CNG (huge, heavy tank required).

Notice in the study intake air heating did two things:

1. Reduce charge air density, therefore torque. Good thing, when trying to keep power down to maintain idle with no throttle.
2. Increase lean burn limit. By the looks of their results, and of the ones posted at top, having it hotter in the cylinder certainly makes sustaining combustion of a lean mixture easier. True, you can heat TOO much, but the correct amount of heating certainly helps with the lean burning, and power reduction required.

I could of course use a tiny amount of supplimental hydrogen to extend the lean burn limits at light load, instead of intake air heating. Medium load should be fine with no hydrogen. The aim would be to use as little as is needed to boost combustion of very lean mixtures.

 

[franzh]

Obanion:
I think that at this point I have tried to present what I think is a reasonable scenario, and that it's now lab time for you. You should locate an engine, install an LPG system on it with enough adjustment to allow for the type of fuel mixture control, bring it up to about 2000 rpm and begin making lean mixture adjustments.

You will quickly see that there is not enough fuel present in sufficient density which would allow the engine to a reasonable low speed that would be acceptable. You will quickly see that as the mixture approaches lean limits, the rpm will drop, and at one point, VERY rapidly quit running.

Your challenge, should you choose to accept it, would be to develop an electronic means of mixture control that would provide the necessary trim levels to operate at the extremely precise levels you need.

Adding H2 MIGHT help some, but I am not sure. That would be an entirely new strategy in your fuel development process. In my experience, H2 in an IC engine is very difficult to manage.

This is the reason the Otto cycle engine is still prevalent, even with its drawbacks of suction throttling losses. If it were even reasonably feasible, I feel we would have seen a little more in the research phase, but there just isn’t so.

Again, not to rain on your ideas, but you asked for input, and that, I have provided.

Regards and best wishes;
Franz

 

[obanion]

Well as you say, if the rpms drop, THATS WHAT I WANT.

My fear is I'll be stuck in a siatuation like this:

1. Above the misfire threshold, but still has enough power to continually accelerate the rpm of the engine.

OR.

2. Misfire, and the engine dies.

If as you say, I can run it lean enough that the rpms will hold steady at 1000RPM, give or take, I'm happy.

I'm fairly certain I have the means of precise mixture control. I have a aftermarket fuel computer with a fully programmable fuel mapping, which can take in RPM, load, intake temp, and throttle position. I plan to use a proportional control valve with under a 1% hysteris.

Yes, I will test this. The startup/idle, and light acceleration test on a car of mine will be the first.

I should just say out with the hydrogen. If I need H2 to make it work, it's not worth it. Rather put the throttle back in then keep 4000psi H2 on board and fueled.

 

[beaverjones]

why take the throttle off anyway?engine will have to work harder to compress full cylinders.

as for the propane injection,theres several turbo'd cars running 20+psi with secondary propane injection for it's anti-knock qualities.works well.