Saar

ideas?

i've come up with a reasonably accurate fuel flow equation in an attempt to make a better MPG calc. now after this project was completed, the idea of using it in combination with an accurate TQ/HP calc to create a BSFC table became a rather interesting idea, though i'm stuck at trying to come up with a consistent way of measuring or estimating torque/HP.

i've looked into some of the PCM code($2E specifically) that deals with how IT estimates torque, but it seems it uses a table that has engine efficiency as an input, which doesn't work for me....

one method i've considered was grabbing MAF readings while measuring torque on a dyno, then scaling torque based on a percentage of the MAF reading seen at lesser throttle openings at appropriate RPMs, but that seems like it would be a bit out of reach for the average vehicle and assuming torque rises in a linear fashion with increasing airmass at a given RPM probably is a false assumption.



so, i'm at a bit of a standstill, no?

EagleMark

Subscribed.

How about whatever a computer software dyno uses? I really have no idea just wanted to subscribe here...

Saar

that would be software based modeling of an engine, which isn't always accurate, at least not with the cheaper software, desktop dyno gave me some rather odd numbers using accurate measurements.

there's another package out there i've played with once, INCREDIBLY detailed, but also very $$$$$$$.

tequilaboy

Load or LV8 already represents a sort of indicated torque. All you need to do is figure out how to scale it into the unit of choice.

Of course this method will neglect inertial effects, combustion efficiency and any air flow or frictional losses, but will be a reasonable indicator.

Other method is to work backwards from the vehicle's acceleration, considering vehicle mass, gearing and driveline friction and inertial losses.

I'd use LV8 and call it a day.

Saar

if i were using a mask that natively used a MAF, i would have LV8 available, but i'm patching it in for fueling only in $A1, so that's a no-go.

tequilaboy

If you have a MAF flow signal and rpm available, you can calculate load. Basically MAF (gm/sec)/rpm * factor.

gm/sec/rpm * 4800 will get you close to an unfiltered LV8.

Saar

but how would i be able to consistently turn the LV8 into a usable lb-ft torque measurement?

my test engine for this should be developing ~180 ft-lb of torque somewhere between 3600 and 4000 RPM. should i look at getting a log of what the MAF says when at WOT in those RPM ranges, calculate LV8 from that, consider the LV8 seen at that RPM to be 100% torque, then scale down torque based on LV8?

tequilaboy

Yes, that should work. You can always adjust the factor accordingly per number of cylinders or so that your 100% torque is represented as a convenient value.

If 100% torque = 180 ft-lb as in your example, then scale your factor to equal 180 at the expected rpm.

Just playing with numbers here, assuming maybe 150 gm/sec at 4,000 rpm. Using the standard factor gives:

150 gm/sec/4000 rpm * 4800 = 180.

Adjust factor as desired to approximate reality.

Saar

hmm..... but without knowing torque at other RPMs, wouldn't that be assuming the engine has the exact same efficiency at 4000 as it does at 2000, or 6000?

EagleMark

Wondering if this will help your calcuation...

http://forum.tunerpro.net/viewtopic....401&highlight=

Saar

not quite, Mark, that was actually the old equation i was using for determining fuel economy, except i believe that is only intended for a V8 in double-fire.

EagleMark

OK. Just trying to help. Your doing math and stuff that's WAY over my head.

The_Punisher454

I use engine analyzer pro, which is MUCH better than desktop dyno (in all its various repackaged forms) and have observed that engines generally have the lowest BSFC at max torque, but will have BSFC numbers higher below max torque and BSFC will increase rapidly at the top end.

For example, on a recent simulatoion of a 427 big block it was showing a bsfc of about .42 or so at max torque but was around .46-.48 at lower rpm's and then climbed up to around .47-.49 at max hp, and then rapidly increased up over .60 as the horsepower curve fell off at the top end.

With the BSFC being on a bell shaped curve it will be difficult to determine actual torque even if you have accurate measurements of Airflow, fuel flow, and AFR.
Incorporating an accelerometer or actual torque sensor may be the way to go.

Saar

right, BSFC being lowest at torque peak is where the whole "an engine is most efficient at peak torque" rule came about... being lower under and above the peak RPM makes sense as well.

torque sensors.... this is a new concept to me.....

The_Punisher454

Apparently there are some transmissions with torque sensors, and I think I saw a tailshaft type solution once a couple of years ago. Probably the same technique used as an electronic torque wrench.

Honestly the only way I can think of to make it work is pressure transducers on the motor mounts with a free floating transmount.

Saar

hmm....

the concept of being able to accurately measure torque on a moving vehicle via a sensor somewhere in the drivetrain is currently magic to me.... i guess i should do some googling on the subject to see what i can come up with.

EagleMark

Some neat math for you Robert, it's in an article I read this morning, about half way down the page on link below. Don't know if it helps here but you would probably appreciate it, in it's complexities. It makes my brain hurt!

http://en.wikipedia.org/wiki/Fuel_injection

Saar

that's actually where i based my new equation from, just modified it slightly to apply to more of the GM OBD1 vehicles, since most of us are running in double fire MPFI.

Craig Moates

Accelerometry will give you good numbers for applied/effective power.

We've got a little gadget that I've been playing with (ADXL345-based) which talks to TPRT5 along with the regular datastream (special add-in). Does a nice job of calculating 60',1/8,1/4,etc info. Add some wind resistance and there it is. Seems to be very consistent and accurate, and there's a ton of imbedded fun to be had.

Nice thing about pulling it off of accelerometry is it gives you what you want instead of having to calculate. F=ma instead of F=m*d2x/dt2 (wheel/shaft sensor) or F=mdv/dt (speedo #). For racing/performance, what are we all trying to optimize? The answer is "a".

Comparing the accelerometry to the MAF readings, RPM, and associated gearing would be pretty slick. Then you could determine losses and such. Wheel slip and/or clutch slip would also be available. If planning some reaction to various trends, you'd want to filter transitions (gear shifts, converter flare, etc).

That motor mount scale deal has been done as well, apparently it worked pretty good (somebody's research project). They had issues with lateral forces and such. There's also a special u-joint or driveshaft coupling available if I recall to measure force.

Who'd have thought? Here we are trying to reconcile PW, MAF, RPM, VSS, AFR, and A. What's next? Hahahaha!

Saar

hmm..... so it's basically a mobile "dyno" using an accelerometer.... similar to all of the G-Tech modules and it's competition?

would it's data be useful in a non-WOT situation? currently, i'm looking for the end-all way of tuning for fuel economy, getting accurate BSFC values seems like it would be the greatest method to do so, since it directly shows HP generated compared to the amount of fuel used... although modifying the conecpt of BSFC to be based on torque generated per fuel used may be a better route.