Drive water injection off of EGT probe data?
 

I was thinking that the best way of measuring the need for water injection is to drive it off of an EGT probe. This way, water is only injected when it offers the benefit of protecting against detonation on a given tune. I'm thinking fail-safe against a heat soaked intercooler or whatever, rather than way of getting more power. Since there is nothing new under the sun, anybody doing this here or know of something like this?

I've not seen this done before.


However, here's my take on it: you want to prevent det/knock/heat BEFORE it takes place, not after. Just like with knock, you want to prevent it from happening, not react to it.

So if your EGT's shoot up, you're already hot and have to cool the engine off. I'd want to prevent it from getting hot in the 1st place.

It is still quote a good idea in my mind as a semi fail-safe.

In for answers from clever people.

From the FAA studies on aircraft, knock develops quickly, but it still takes hundreds of revolutions for it to go from warmer than typical to all out knock.

My thinking was more that the EGT rises in response to boost, more so with more boost. If that were all there were to it, boost linked injection would be as good as is possible. Making up some numbers to make it easier to explain:



- Lets say 950 F is the max EGT with boost disabled, on a hot day.
- Let's say 1125 F is detonation.
- Let's say typically the engine can go to 10 psi of boost with EGT at 1050 without water injection.
- Let's say the EGT that indicates the onset of pre-ignition is 1075 F measured at an EGT probe just at the turbine flange (This is the only number that didn't come out of my ass, it came out of Corky's :D ).


See where I'm going with this? We could set such a system to start injecting water at 1025 F, ramp up to maximum injection at say 1050 F, and turn on an indicator light while doing max injection above that. Since it is reading EGT and has an indicator light, we could even make it blink in anger at 1075 F if we wanted. Of course, the EGT wouldn't keep rising with boost, since boost would cause a rise that makes the system inject water which lower the EGT. The feedback could dramatically lower the amount of water consumed, but yet increase it at the critical moments that might normally blow the engine on a decent tune under bad conditions.

Alternately, the control could be set to a lower value to extract more HP, or higher to be a backup fail safe.

The effect of such a system would be to inject water in response to rising likelihood of detonation, rather than in response to rising boost, so it would inject less water at higher boost under cooler conditions and more water at lower pressure under hot conditions.

TL-DR: I'm basically saying it is the cyl temps we want to control. Generally, when boost rises, so do temps, but the starting point and rate of rise can differ depending on other factors. I'm saying the EGT is directly proportionate to cyl temps, which are directly proportionate to knock probability for a particular engine. So boost linked injection is like open loop, and EGT linked injection is like closed loop. With a smart program, closed loop is always better.

EGTs will drop when there is detonation, thats one of the reasons it is so useful, but only if you know what to look for

Isn't it too late by the time it gets that severe?

Possibly.

But anyway. I dont think its necessary to reinvent the wheel. Theres a pretty good relationship between boost and the need for anti-detonate.
As long as it keeps flowing, that water is going to do an astounding job at preventing detonation.

I was kinda wondering that too: is the point of this suggestion to inject less and save on meth and/or water? lol

Cause otherwise "the wheel" is just fine and doesn't need re-inventing IMO

I am assuming the point was the efficiency of the system. Not wasting water/meth when you don't need it. If you can make it work I think it is a good idea. There is always room for improvement.

you could use megasquirt to trigger water injection above a set temp say(120F) from the IAT sensor on the cold side of the intercooler. then you have the ability to tell it above what rpm threshold and map preassure. just a thought