engine tuning vs. ambient conditions

[Juan Piñeiro]

Hi there.

The engine guy in our team and I are working on a project to optimize the engine tuning of our FSAE car. This project will go from the generation of the best possible fuel and spark tables to get the desired torque, power and noise behavior from the engine (using a dyno) and also finding some direct correlation between ambient conditions (humidity, barometric pressure, dry bulb temperature, etc) and those tables.

In other words, we want to develop you can say... some sort of "algorithm" to fuel and spark tune the engine for "any" ambient condition.

What would be a good starting point for you guys for this ambient conditions/tuning correlation? What books or papers would you recommend?

Thanks in advance

Juan Piñeiro
Drivetrain & Brakes leader
FSAE LUZ (Venezuela)

[Grant Mahler]

Originally posted by Juan Piñeiro:
In other words, we want to develop you can say... some sort of "algorithm" to fuel and spark tune the engine for "any" ambient condition.

What would be a good starting point for you guys for this ambient conditions/tuning correlation? What books or papers would you recommend?



All of the Megasquirt (http://www.msefi.com) source code is published. You can get all of their algorithms. Whether you think they are worth anything is up to you, but it gets you some idea of where to go.

I don't know of any other company that publishes the algorithms.

Several teams build and write their own ECUs. Cornell is the first that pops to mind.

[murpia]

Firstly, most ECUs do not use fuel & spark 'algorithms' but look-up tables. Understanding that distinction and the pros and cons is an important first step.

Secondly, most ECU control strategies are limited by the available sensor technology and cost, when it comes to compensating for 'any' ambient conditions.

Anyway, Megasquirt is a very good place to start.

Regards, Ian

[Juan Piñeiro]

Thanks for the answers, I'll try megasquirt to see what I can find.

I'm sorry if I wasn't very clear about what we wanted, it's not that we want to program our own ECU or whatsoever, we're not at that level yet.

What we really want to do is to find some correlation between ambient conditions and load lines (fuel and spark) in an experimental way, using dimensionless parameters or something like that.

In other words, we want to collect some data that can tell us something about that correlation in order to optimize engine tuning under almost any ambient condition.

Does such a correlation or function already exist? maybe some standard or correction factor?

[B Lewis @ PE Engine Management]

Hi Juan,

There are emperical methods for getting baseline fuel numbers based on ambient conditions, however, "calculating" the effect of ambient conditions on timing is much more difficult. OEMs still spend 100s of hours experimentally tuning engines despite having very sophisticated analysis tools.

[Pete M]

For fuel, ambient corrections are fairly easy to make up. You get fairly close with ideal gas assumptions. We usually then fine tune the temperature one (with a turbo it's easy, manifold temp rockets up when you cover the intercooler. Of course, temp compensation is probably far more important for us).

Ignition is harder, and the OEMs use a lot of expensive equipment and spend a lot of time altitude testing etc, to get it right. But with some creativity and time you can give yourself a few data points.

The question you need to ask is how much variation do you expect your car to need to handle? Altitude is a bigger effect than some people realise. Cali in 06 was 95-96 kPa most of the comp rather than the sea level 101 kPa (5-6% density). Temp you definitely need to correct for across some reasonable band. If you're serious, don't neglect the others such as engine temp comp, fuel pressure comp, fuel temperature comp... well, the list is rather long.

[Juan Piñeiro]

Thanks for all the replies I really found them useful for our project. It's very important to us because in this year's competition the engine of our car was not properly tuned due to the lack of experience in that area.

This year we're looking into some dyno testing and data acquisition http://fsae.com/groupee_common/emoticons/icon_smile.gif thanks again for all the help.