Dynamic Cylinder Deactivation

[christofoo]

Originally Posted by t vago

So, we have a set amount of produced work, now per firing cycle instead of per cylinder. We have a set amount of pumping work being consumed per firing cycle, instead of per cylinder. We produce a set amount of available work per firing cycle, instead of per cylinder.

We shut off one cylinder via DCD. The remaining cylinders are still producing work. You pointed out that the remaining cylinders are producing less work per firing cycle, than before.

So, we need to increase the amount of fuel to the remaining cylinders, so they can provide the amount of available work as before, while also covering the slack of the deadbeat cylinder caused by DCD.

Hm... The other cylinders are already operating at stoich, right? So we must also increase the amount of oxygen going to the running cylinders! Okay, so we lowered intake manifold vacuum by a bit. Good deal! We just lowered the pumping work for all of the cylinders, right?

Waitaminute... We just increased the amount of fuel to the remaining cylinders. We did this to cover the deadbeat cylinder, which is still consuming pumping work. The pumping work per cylinder may be less than before, since we just lowered intake manifold vacuum, but we also increased the work produced by the remaining working cylinders.

Right?

Very good, hang on, we're starting to converge now.

Which is greater, the increase in fuel consumption or reduced pumping loss? Let's back up and set the problem up to work out the same way you analyzed lean burn. Start by taking the fuel that would have gone to cylinder 1, and distribute that fuel to the other 3 cylinders. Cylinder 1 is cut, now we're consuming the same amount of fuel per cycle with 3 cylinders that we would have consumed with 4.

But, as you pointed out, the fuel per cylinder has gone up, so manifold vacuum must go down to get more oxygen into each cylinder to maintain stoich AFR.

Agreed?

[t vago]

Originally Posted by christofoo

Very good, hang on, we're starting to converge now.

Which is greater, the increase in fuel consumption or reduced pumping loss? Let's back up and set the problem up to work out the same way you analyzed lean burn. Start by taking the fuel that would have gone to cylinder 1, and distribute that fuel to the other 3 cylinders. Cylinder 1 is cut, now we're consuming the same amount of fuel per cycle with 3 cylinders that we would have consumed with 4.

But, as you pointed out, the fuel per cylinder has gone up, so manifold vacuum must go down to get more oxygen into each cylinder to maintain stoich AFR.

Agreed?

Right.

However, if we assume that we keep available work per firing cycle constant, we have to increase the available work per firing cylinder to cover the lost available work from the deadbeat cylinder. This requires that the firing cylinders' produced work must increase, which means more fuel being consumed by the firing cylinders in the firing cycle.

We must also cover the deadbeat cylinder's pumping work, which means that we must increase the firing cylinder's produced work that much more. This means even more fuel must be consumed by the firing cylinders.

All these things will lower intake manifold vacuum in order to provide enough oxygen to the firing cylinders, which will in turn lower the pumping work being consumed by all of the cylinders. However, are you really reducing the per-firing-cycle fuel being consumed?

[christofoo]

Originally Posted by t vago

...
However, if we assume that we keep available work per firing cycle constant, we have to increase the available work per firing cylinder to cover the lost available work from the deadbeat cylinder. This requires that the firing cylinders' produced work must increase, which means more fuel being consumed by the firing cylinders in the firing cycle.
...

I don't think we've reached the same page just yet. We're still stuck on the step: going from 4 cylinders firing to 3 cylinders firing.

With 4 cylinders firing, the available work and pumping losses are exactly met by the produced work.

We're not going to change fuel input yet (EDIT: fuel input per cycle), we're going to redirect fuel from 1 deactivated cylinder to 3 active cylinders.

Pumping losses have not changed yet. We had 4 cylinders before deactivation and there are still 4 cylinders after deactivation that are all experiencing the same pumping loss. Inlet and outlet pressures have not changed yet. Total cycle pumping loss has not changed yet.

Produced work is dictated by a set amount of combustion components, which are not going to change because the fuel per cycle hasn't changed yet. The AFR is broken right now but we're going to fix that in the next step.

The next step is to adjust manifold vacuum to correct the AFR.

[t vago]

Alrighty, then. I've demonstated lean burn. Now, you demonstrate how DCD is supposed to work.

Oh, and if DCD is supposed to save fuel, then why not just permanently deactivate a cylinder, and vent its exhaust directly to the atmosphere? After all, there should be no fuel vapors as there is no fuel going to that cylinder.

[christofoo]

Originally Posted by t vago

Alrighty, then. I've demonstated lean burn. Now, you demonstrate how DCD is supposed to work.
...

? Are you agreeing with me now? Do you disagree with the problem setup in #253? Does it require clarification?

(I thought I set it up the same way you setup your explanation of lean-burn, which, by the way, was excellent and I'm very glad for the time you put into it.)

[t vago]

Originally Posted by christofoo

? Are you agreeing with me now?

Nope.

Originally Posted by christofoo

Do you disagree with the problem setup in #253? Does it require clarification?

Nope.

Originally Posted by christofoo

(I thought I set it up the same way you setup your explanation of lean-burn, which, by the way, was excellent and I'm very glad for the time you put into it.)

Well, I'm glad you like it. Now, you can put in the same amount of time, trying to show how to improve fuel economy by placing an intentional parasitic drag on an internal combustion engine.

[christofoo]

Okay, just didn't understand that comment.

So, we finished #253 after deactivating 1 cylinder, which was done with no change to pumping loss per cycle, no change to fuel injected per cycle, but with an AFR problem.

To rectify the AFR problem and get back to stoichiometric, manifold vacuum must be decreased to increase the mass of air intake in each cylinder.

DCD is now in the same place we were in #243 with lean-burn and the rest of the argument follows. EDIT: to clarify, no change in the number of cylinders experiencing pumping loss, just a change in the manifold vacuum.

[t vago]

Okay. You somehow proved it worked... somehow. Wave a magic wand, and viola!

Now, do what the major auto manufacturers failed to do, what heihetech failed to do, and what everyone else who had this bright idea (of improving fuel economy via placing a de facto air pump in the engine powertrain) failed to do, and make it work.

[christofoo]

Originally Posted by christofoo

...
DCD is now in the same place we were in #243 with lean-burn and the rest of the argument follows. EDIT: to clarify, no change in the number of cylinders experiencing pumping loss, just a change in the manifold vacuum.

Maybe that wasn't as clear as it could be, probably you figured out what I meant, but I can say this more precisely, and I didn't mean to imply that DCD burns lean, which was the conclusion of #243.

What I meant was:

Fuel input has not changed since we started DCD. Produced work has not changed. The number of cylinders experiencing pumping loss has not changed. But manifold pressure has gone up, manifold vacuum has gone down, pumping loss has gone down.

And this means available work has gone up.

[christofoo]

Originally Posted by t vago

Okay. You somehow proved it worked... somehow. Wave a magic wand, and viola!

Now, do what the major auto manufacturers failed to do, what heihetech failed to do, and what everyone else who had this bright idea (of improving fuel economy via placing a de facto air pump in the engine powertrain) failed to do, and make it work.

(Doh, typing while you were.)

I agree implementation has hurdles, I'm not sure it's worth testing at all. Sometimes manufacturers have good reasons. The explanation could be emissions - meaning cat temperature, IMHO - and it could be a big deal. Sometimes manufacturers don't seem to have much of a reason... smooth wheel covers, grille blocks...

I do think I owe you an apology, I'll get back to that in a minute.