Accurate Trip MPG Calculation

[RobertISaar]

me, being the perfectionist that i am, i needed to find a way to come up with perfectly accurate MPG numbers for any trip.

now, i've been able to make accurate instant numbers before using an equation that's been posted a few times, but when averaging that out over a log, it comes out quite optimistic, so i came up with a better method.

now, the faster your datastream is, the better, so 160baud is pretty much not going to benefit much from this since their conditions change so much between samples.

what i do is export a CSV log from a tunerpro XDL file with the following items in it:

vehicle speed
fuel flowrate in gallons/hr

now to allow those two values to show up correctly(at least in the current version of tunerpro), you need to also export the values that fuel flowrate is created from. now it's generated by fuel flowrate in lb/hr, but the lb/hr value is generated via BPW and RPM, so a total list of items exported are:

MPH
flowrate in G/Hr
flowrate in Lb/Hr
BPW
RPM

you can export more, if you feel a need to, but not necessary for the MPG calc.

so, now that you have a CSV file filled with fun numbers, you need to create a few values:

log duration
average fuel flow in gal/hr
total fuel used
average speed
total distance traveled
and finally trip economy

if you've messed with spreadsheets before, this will be fairly simple, but here is how to create each(after scrolling all the way to the bottom of the document):

log duration
this is always the first item i see and is in column B(as Time, Seconds). it's displayed in seconds, so make a value underneath all of the nice numbers and take the last value and have a cell operation of "=B23554/3600". B23554 being the last cell number for this example, now you have the length of the log in hours.

average fuel flow
column E for this example. "=AVERAGE(E4:E23554)" now you have a averaged fuel flow over the entire log. since fuel flow takes RPM into account, cells with the engine not running won't effect the calculation(or in the case of RPM values that can't be 0, very little and to the point of where you could remove them beforehand). displayed in G/Hr.

Total fuel used
this uses the previously two generated values, so no column to worry about. "=C23558*C23556". now those cells are where i have log duration and average fuel flow stored, though yours will vary significantly. just multiply the two together and store them in a different cell. fuel used is displayed as gallons.

average speed
just like average fuel flow, just in a different column. for me, column G. "=AVERAGE(G4:G23554)". now we have average speed over the entire log. displayed in MPH

total distance travelled
just like total fuel used. multiply average speed with log time. "=C23561*C23556" and now we have miles traveled.

finally, we take miles traveled and fuel used to make miles traveled per gallon used(MPG)
"=C23562/C23559", with '562 being distance traveled and '559 being fuel used.

congratulations, you should now have something looking similar to this:



http://i.imgur.com/zagQj.png

keep in mind, i do this all in openoffice, so MS and others may be different.

[gregs78cam]

I would love to use this through a whole tank of gas to see just accurate the BPW vs RPM conversion is. Ya know, just how accurate the fuel flow that the ECM sees is.

[RobertISaar]

i've been tracking it with the MC, and it's very close in my application. within a few percent.

keep in mind it does not account for any async fuel, though if you're driving for fuel economy, you shouldn't run into that too much.

[EagleMark]

so 160baud is pretty much not going to benefit much from this since their conditions change so much between samples.

You saw me coming didn't you? I do have a P4 conversion half done on sons truck and an LT1 too you know...

How accurate is the MPG history tables you make? Seems you like this, how do they compare or how do they compare to actual gallons used at pump?

[RobertISaar]

using this log as an example, TP calc'd it out to be like 47MPG, which is optimistic to say the least.

using this method, you see 34.8MPG, which is really believable for this car when i'm driving it nicely(and have the benefit of a lot of downhill coasting).

luckily, this car doesn't have a hole in it's fuel tank, so i can do full fills/empties, however, i do have a leaking fuel neck, so when i do fuel up, some of it ends up on the ground. i'll have to attempt to log an entire tank without a single moment of the engine running not being logged.

a little earlier in the year(when it was significantly colder and i let it warm up more), i was still seeing 28MPG on similar routes using the tank empty method, though with the fuel neck issue, i get slightly better than what that comes out to, but since i'm still paying for it, i don't attempt to compensate for it.

also, 160 baud will work..... but don't expect it to be AS accurate, since a lot of things can go on between a sample with it.

[EagleMark]

also, 160 baud will work..... but don't expect it to be AS accurate, since a lot of things can go on between a sample with it.

I was going to build a MPG history table for the $42 since it is a very nice advanced type file, and I still own a stock factory 1227747 1990 Suburban I've had for about 10 years. Do you think it would be worth the effort?

I'm on to 16197427 within the month with tuning the kids IH 345, Isky cam bored, balenced, exhaust ported new engine! Woo Hoo! I finally have one to play with. I'm wore out on the LT1... and all the 1227747 Off Road Conversions I take care of around here.

[RobertISaar]

it's not difficult to setup a MPG estimate. the way i mentioned...

in fact, we've discussed it recently, it's just that this method is more accurate for a trip rather than letting TP's histogram come up with an average.

[EagleMark]

it's not difficult to setup a MPG estimate. the way i mentioned...

in fact, we've discussed it recently, it's just that this method is more accurate for a trip rather than letting TP's histogram come up with an average.

That was with the LT1 and comes up fairly accurate!

What I was asking is do you think it's worth the effort on the 160 baud 1227747?

[RobertISaar]

yeah, i mean it only takes a few minutes to setup the values in tunerpro, since now BPW has been found for the 7747.

worst-case, it's a learning experience.

[RobertISaar]

another example from earlier today:


http://i.imgur.com/8Ywj5.png

this was from a cold start (33*F), i removed the first 27 samples since those are what were logged between my hitting record and actually cranking the engine. i changed one equation at the end to account for the first 2.557 seconds not being in the log, but otherwise, the same process.


http://i.imgur.com/l315E.png

and i ended up at 29.6MPG, that's start to end, cranking fuel included. not bad considering the horrible winter blend fuel we have for now, combined with a bad alignment, increased air resistance due to massive amounts of wind and cold air and i have all emissions related stuff disabled so i can track down a random knock and do a final VE tune.

[EagleMark]

29.5 MPG is nothing to complain about at slow speeds in winter!

I did do the MPG for 1227747 history table and seems to be fairly accurate compared to know MPG which at best is 12 below yours...

[RobertISaar]

191 cubic inches TEND to be fairly economical. and slow speeds isn't necessarily true. most of my route is at 55MPH with the exception of 3 towns i have to go through to avoid going through any unnecessary slow turns/stops since momentum is key to awesome fuel economy. it doesn't help that elevation doesn't remain constant at all around here, it's all uphill, downhill, 1/4 mile flat section, 55MPH curve, go uphill, repeat.

and do be sure to try both trip MPG calc methods, i've found the tunerpro history version to be quite optimistic for me(not 100% certain why, but i have a theory). doing it via spreadsheet is spot on with my last couple of tanks.

[EagleMark]

Miles divided by gallons works for me. But I put it in to see if there's a sweet spot between 60 and 75 MPH... this Suburban is a brick with 3:73 gears...

That is a sweet spreadsheet project you did though!

[RobertISaar]

i'm running 3.33 with this car(and 27.7" tires), but when the 91GP SE gets on the road, it will have 3.43(and 26.2" tires), though i may jump up to 3.69 or 3.73 with it. DOHC like to spin. if i manage to go 6 speed manual with it(there is a trans that has the perfect bellhousing pattern), then i'll have a 3.55FDR and a .62:1 6th gear.

anyways, the theory of the spreadsheet, in my eyes is flawless, since it simply takes gallons injected and divides with miles traveled to come up with MPG. obviously, a leaking/sticking/failed injector would skew these, but a properly running engine will work with it.