Correlation between predicted and real-world fe?

[Frank Lee]

Thought of this when commenting on the Zoltar Miracle Trike:

The Ecomodder fe prediction spreadsheet is a wonderful tool- takes a lot of onerous calculations away! But one limitation of it is, the values given are valid for steady-state, super calm and perfect conditions. As an example, I plug in the values for my Sport Coupe and get what I think are realistic predicted fe values at the steady state speeds, but tank-to-tank proven fe never matches that (nor should it) due to the real world getting in the way.

So I looked at it a bit and it seems in the case of my Sport Coupe I have to knock 10-15% off the predicted fe to get the real world fe.

That is with normal attentive conservative driving, not extreme P&G, of course.

I wonder if that 10-15% correction factor is valid for everyone else's predicted/real fe too?

[MetroMPG]

Another shortcoming of the calculator is that it doesn't take brake specific fuel consumption (BSFC) into account.

For people who regularly do long-haul drives at a steady speed, it'd be interesting to know how close in the real world they get to the model though.

[RobertSmalls]

On my 10 mi commute, I find that my instantaneous fuel economy during steady state cruising once warmed up, is about 30-40% higher than my trip average. Commuting represents a large chunk of my fuel spent, so my tank average is well below my steady state cruising mpg.

However, the calculator is within 10% of my observed steady state cruising mpg, which tells me my estimates regarding actual BSFC, Crr, and CdA were pretty close.

[roflwaffle]

Originally Posted by MetroMPG

Another shortcoming of the calculator is that it doesn't take brake specific fuel consumption (BSFC) into account.

Sure it does. Engine efficiency = BSFC. A gas engine operating at ~100% efficiency would use ~80+g/kWh, so if an engine is at ~300g/kWh it's at ~27-28% efficiency. The 2nd gen Prius at a minimum of 220-230g/kWh would be at ~37% in terms of peak efficiency, which is coincidentally (not really) what it's at. Toss in ~90-95% transmission efficiency for a locked-up auto or a manual transmission and bob's your uncle. I think it could diverge the most for an automatic trans that's always unlocked, but those aren't very common. Also the actual gasoline weight/energy content could throw things off since there's no standard for a BSFC test AFAIK.

[luvit]

frank, for a living i analyze data everyday - but not data like this calculator.
i respect the error margin that you determined.

have you determined this on all 3 of your listed vehicles?
do you still have 3 or more?

this is something the the calculator could contribute, for the users who may miss this thread. the calculator would show it's own calculations, but recommend #% be deducted without having to rewrite any code.

[Frank Lee]

I only did the numbers for the Sport Coupe.

Jah, "Blue" has an old school a/t w/o lockup

Just wondering if anyone else thought of this and if their logs show that their average fe is 10-15% off "perfect". If it is true for enough people, perhaps it would be a good correction factor for the fe tool users.

[cfg83]

Hello -

Maybe you could apply a set of theoretical "trips". Imagine a track with hills and valleys, stoplights, and traffic jams.

One thing I did notice is that the calculator does not take into account what gear you are in. However, that is already discussed in the thread that is dedicated to the calculator :

Online tool: shows aerodynamic drag, rolling resistance, power required & est. MPG

In terms of correction factor, it must be a range based on context. Urban vs rural, snowy winters vs mild winters, windy regions vs calm regions, etc ...

I look at the calculator as the "MPG Limit" of my vehicle. I think that it does allow you to do a "what if I lower my Cd by X amount" and see what happens to your MPG.

CarloSW2

[Frank Lee]

Yeah, there's probably too many variables to make a blanket statement about it. I just thought it would be nice to punch in the values, knock 10-15% off, and say "yup, that's about what I'll get" especially for stuff that has no EPA figures, like old stuff or theoretical stuff.

Maybe it would work for each individual, though. Like, I know my environment, and if I show that all my vehicles return 90% of predicted fe for me, in my area, then I can apply that to the old/theoretical for a decent guesstimate.

[roflwaffle]

I'm totally BS'ing here, but it would be neat to have the mpguino dump stats to a computer. Combine that w/ something like g-maps pedometer and it's elevation changes, and we could construct a multi-dimensional (including engine operating temperature too for instance) BSFC map of the current engine. Once everything is nice and calibrated, the driver could input a route and a program could give on the fly suggestions regarding efficient driving.

[RobertSmalls]

It might be easy to have the calculator spit out a weighted average of your fuel consumption at 35, 45, 55, and 65mph, with a few braking, idling, and hill-climbing events thrown in, and maybe a warm-up period. But to get that weighted average to reflect every individual's actual driving habits? Forget it. Different drivers have different rates of acceleration, cruising speed, and methods of deceleration (DWB / EOC / DFCO down to x mph, regenerative braking, or "leave it in D").