Here's one for the performance gurus...
In perusing one of the many NACA WWII wartime research papers that are now available online, I noticed an unexpected trend in knock limited performance data. With knock being held more or less constant at the incipient level, and increasing fuel/air ratio, after the expected increase in knock-limited power with enrichment, the curve inflects and the knock-limited power decreases. The data also show, this was more than a loss of thermal efficiency due to over-enrichment. It was actually necessary to reduce charge air flow with increasing enrichment, to maintain the same level of knock. Here is a figure showing this tendency.


Here is the only comment from the paper that obliquely addresses this phenomenon:

This phenomenon can be seen in many of the other figures 1-8.
The full paper can be downloaded here: NASA website
I'd be interested in explanations for this phenomenon.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

The follow-up report shows the same phenomenon in several graphs.
Follow-up report

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

In case the linked images above aren't working, here they are again.



I finally found a paper with a possible explanation. Here is the explanation.

And here is an example of their data.

I don't think it is conclusive. Essentially they're saying that, while the bulk combustion is starting to lose some potency due to the excessively rich mixture, conditions are still agressive enough to cause end-gas knock.
Here is a link to the complete paper.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

hemi, I could not get the followup report. I'll try again later. However, I've studied many of these wartime reports. They are beyond fascinating.

What is being shown is often not exactly obvious because they are changing conditions to maintain some parameter constant or adjusting for a maximum.
You already know what follows:
Enrichment increases power by ensuring the maximum reaction of available oxygen.
Enrichment can further increase power by suppressing detonation: in the case of a supercharged engine, allowing more boost; in case of an NA engine, allowing the engine to live with a higher CR.
Enrichment suppresses detonation by at least two mechanisms: by acting as an internal coolant; by acting as a diluent in that overly excessive fuel actually slows the reaction.
When the second effect begins to come into play, power is reduced. In a test where maximum power is being sought, the operator's adjustment of boost to produce a certain level of power will run into detonation at a lower realized power level as enrichment is increased. It will appear that increasing enrichment produces increased detonation where actually it is only reducing power.
When water is substituted for fuel enrichment, the cooling effect is much greater and the allowable boost level goes up quite a bit. But, water is a diluent too so you might see some of the same confusing indications of adding water actually APPEARING to promote detonation. However, they seemed to almost never take the water proportion beyond 60% or so. Also hard to decipher is the effect of adding Methanol. It is not as good an internal coolant as water, but in part it is acting as a fuel substitution and it is also promoting the complete evaporation of water in the manifold, increasing the density of charge at a given boost level.

Another VERY big distortion in these data is the fact that engine cooling and supercharging are being provided by outside power sources. These loads are big influences on how the engine will perform in the field and are sensitive to the parameters being tested. I think they found that the turbocharged engines performed better than expected compared to supercharged engines given the test data.

Anyway, that was what I got out of these reports. What do you think??

The methanol also stops the water freezing at altitude which is maybe its original purpose.

If to much is added, a power gain may be had by actually cutting main fuel a little as the water/methanol blend is added over a certain level..

Regards
Pat
See FAQ731-376: Eng-Tips.com Forum Policies for tips on use of eng-tips by professional engineers &
http://eng-tips.com/market.cfm
for site rules

I was really just focussing on the the "knock-limited" power aspect. I understand that beyond a certain enrichment (or dilution, or internal cooling, as the case may be), power goes down, knocking or not. What caught my eye was knock-limited power decreasing with enrichment, i.e., with knock being held constant. Prior to encountering these papers, my naiive assumption was that, as a particular engine is manouevered along its knock-limited power vs enrichment trajectory, power would monotonically increase until both power and knock fall off precipitously; i.e, at the point beyond peak power, the limitation is no longer knock but combustion efficiency. To be sure, many such plots in these papers indicate that exact outcome, i.e. the knock-limited power curve is still trending upward with enrichment, at the furthest extent of the data. It is the contrary data that puzzle me and are the cause of this thread.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

Ok. My take was whatever power is being made will be knock limited because the operator will force that to happen. So, with the engine at its absolute maximum knock limited power under some particular state of enrichment, further enrichment will not precipitate knock, but will reduce power. Then the operator will increase boost until knock is induced. The power reading will now be recorded as the knock-limited power under the increased enrichment and it will be lower than the power recorded with the previous enrichment setting. It will look on the graph like increasing enrichment induced knock while that is not the case. Unless I am misreading the procedures.

It's getting late for me now, so I'll put that in my pipe and smoke it, and get back to you. Thanks for the feedback.

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz