Throttle-less intake

[gruntguru]

In automotive applications the throttle controls engine load (torque) - not rpm. RPM is controlled by the transmission.

Skipping injection is a good strategy but the WBO2 will read lean due to the unburned O2 from skipped cylinders.

You need to map injection quantity to engine speed. Target a rich mixture (lambda approx 0.9) at full load and lean (1.1 - 1.2) at part load. That alone will reduce the output by about 20%. Depending on engine design you may be able to go a little leaner. Note that lean mixtures will require more spark advance. The WBO2 will give useful readings throughout this range.

Once you have mapped the entire load - speed range (from 80% to 100% load) you can produce the 0% - 80% region by mapping your skipping strategy. Injection quantity and ignition timing will be the same as 80% load at each speed.

[langwadt]

In automotive applications the throttle controls engine load (torque) - not rpm. RPM is controlled by the transmission.

Skipping injection is a good strategy but the WBO2 will read lean due to the unburned O2 from skipped cylinders.

You need to map injection quantity to engine speed. Target a rich mixture (lambda approx 0.9) at full load and lean (1.1 - 1.2) at part load. That alone will reduce the output by about 20%. Depending on engine design you may be able to go a little leaner. Note that lean mixtures will require more spark advance. The WBO2 will give useful readings throughout this range.

Once you have mapped the entire load - speed range (from 80% to 100% load) you can produce the 0% - 80% region by mapping your skipping strategy. Injection quantity and ignition timing will be the same as 80% load at each speed.

with out direct injection I foresee the problem that the fuel actually entering the cylinders per intake won't be the same
as mapped when you start skipping injections

and a cat won't work

[gruntguru]

In automotive applications the throttle controls engine load (torque) - not rpm. RPM is controlled by the transmission.

Skipping injection is a good strategy but the WBO2 will read lean due to the unburned O2 from skipped cylinders.

You need to map injection quantity to engine speed. Target a rich mixture (lambda approx 0.9) at full load and lean (1.1 - 1.2) at part load. That alone will reduce the output by about 20%. Depending on engine design you may be able to go a little leaner. Note that lean mixtures will require more spark advance. The WBO2 will give useful readings throughout this range.

Once you have mapped the entire load - speed range (from 80% to 100% load) you can produce the 0% - 80% region by mapping your skipping strategy. Injection quantity and ignition timing will be the same as 80% load at each speed.

with out direct injection I foresee the problem that the fuel actually entering the cylinders per intake won't be the same
as mapped when you start skipping injections

Why? Flow on the intake side of the engine is little affected by whether the charge is fired or not - perhaps to a larger extent when radical valve timing (large overlap) is employed.

and a cat won't work

An oxidation cat will. Reduction cats won't work at lean mixtures either and the whole thrust of throttle-less intake is load control via enleanment.

In any event, I didn't get the impression that pp51 is developing an engine to meet emission regulations.

[langwadt]

In automotive applications the throttle controls engine load (torque) - not rpm. RPM is controlled by the transmission.

Skipping injection is a good strategy but the WBO2 will read lean due to the unburned O2 from skipped cylinders.

You need to map injection quantity to engine speed. Target a rich mixture (lambda approx 0.9) at full load and lean (1.1 - 1.2) at part load. That alone will reduce the output by about 20%. Depending on engine design you may be able to go a little leaner. Note that lean mixtures will require more spark advance. The WBO2 will give useful readings throughout this range.

Once you have mapped the entire load - speed range (from 80% to 100% load) you can produce the 0% - 80% region by mapping your skipping strategy. Injection quantity and ignition timing will be the same as 80% load at each speed.

with out direct injection I foresee the problem that the fuel actually entering the cylinders per intake won't be the same
as mapped when you start skipping injections

Why? Flow on the intake side of the engine is little affected by whether the charge is fired or not - perhaps to a larger extent when radical valve timing (large overlap) is employed.

i just have the hunch that the fuel is lagging the air to some extend so that each intake is not a separate event, so after a skipped injection where the intake is "sucked dry" the next injection needs to be bigger to compensate

[pp51]

Cheers for the information, guys!

I need to clear some things, though: gruntguru was right, emissions are pretty much last on our list right now. Another issue is that our first round of tests and initial fire-up of the engine will be with no load at all and that the engine is port-injected.

I am far from an expert in engines (software engineer by training), so, since the firsts tests will be at very low load, I am thinking to target a theoretical lambda of 1 (knowing mass of airflow into the engine and injector size) all over the map and work on the algorithm that skips the injection events. Then, as we start putting load on the engine, the target AFR map can be refined to force the engine go lean in certain conditions.

Regarding the algorithm for skipping injection events, are there any pitfalls? I am thinking to first calculate a percentage of events that need to be skipped and then distribute those evenly.

[gruntguru]

langwadt. You are right. A throttle-less engine relying on skipped injection to regulate load needs a fully sequential injection system with end-of-injection timed to occur with intake valve open. This will ensure that the entire injection event is consumed by that particular cycle of that cylinder.

pp51. How do you plan to control the speed of your engine on initial startup? Optimising the fueling and timing can only be done properly on a dynamometer and in a non-skipping mode.

The skipping strategy should aim to be (reasonably) equally spaced and be shared equally by all cylinders to maintain even heat distribution. There are probably some algorithms getting around - skipping (spark or fuel) is extensively used for "soft" rpm limiting.

[riff_raff]

Almost all current SI auto engines use cam phasing mechanisms to vary valve timing. These phasing mechanisms have become sophisticated enough that they can provide some intake throttling functions. They also are very effective at providing EGR functions, which helps with emissions control.