My fuel economy master equation

[alpha754293]

New picture uploaded. The results are in. I've now tested my FFH from 90-120 km/h in terms of fuel economy. There's an interesting observation that I noted while I was doing the testing. (Wow...can I sound ANYMORE like a geek??? lol....yayyy scientific method) In any case, the master equation is up for my FFH.

 

So a little bit of background (this was actually continued from another thread) - there's a section of ON-401 just outside of Windsor (where I live) between Dougall Parkway and French Line which is a good place for me to test fuel economy because it's flat and straight. The total distance works out to be about 16.4 miles with over I think like two very minor changes in elevation (small "humps" like an overpass).

 

I originally tested it a couple of months ago at 120 km/h (74 mph) on the Eastbound (a.k.a. "outbound") run and got 34.3 mpg, but on the Westbound return run (a.k.a. "inbound") I slowed down to 110 km/h (69 mph) and got 37.3 mpg. Last night I ran it again, but slowing down to 100 km/h (62 mph) on the outbound run and got 40.6 mpg, and then slowing it further still to 90 km/h (55 mph) on the return and got 46.0 mpg.

 

Ran the polynomial regression on it and this generated this master equation:

 

y= 0.006x^2-1.644x+145.27

 

where x is speed in km/h

and y is fuel economy in mpg

 

with an R^2 value of 0.99785 (which is an INCREDIBLY accurate regression).

 

And this is before reflashing/upgrading/reprogramming the powertrain control module (PCM) with the new update. There are a couple of interesting observations that I note in the picture, so take a look.

 

Comments are more than welcome.

[hybridbear]

Did you also record the SOC at the beginning and end of testing? That makes a huge difference since your tests cover such a short distance. From my experience you need at least 40 miles on the freeway at a fixed speed to get accurate results because of the huge penalty in the first few miles of getting the SOC up to the ICE High level.

[acdii]

Did you also record the SOC at the beginning and end of testing? That makes a huge difference since your tests cover such a short distance. From my experience you need at least 40 miles on the freeway at a fixed speed to get accurate results because of the huge penalty in the first few miles of getting the SOC up to the ICE High level.

This is true, took me about 6 miles before I got over 40 on the instant, with AC on.

[milleron]

Is the utility of the regression supposed to lie in predicting fuel economy on that stretch of road at different speeds?

 

It doesn't seem to give very realistic values when I try it at slower or faster speeds.

[alpha754293]

This is true, took me about 6 miles before I got over 40 on the instant, with AC on.

 

There's actually quite a bit of a run-up before the on-ramp. Dougall (the street) actually has a posted 50 km/h limit, which I start out at between 60-65 km/h, and then it becomes a 60 km/h limit, where I speed up to 70-75 km/h, and then it becomes Dougall Parkway (limit: 80 km/h) where I go up to 90 km/h, and then finally 100 km/h (about maybe a km before the on-ramp).

 

So to answer the question, no, because the firmware for the OpenXCPlatform CAN bus reader isn't available for the FFHs yet and visual inspection of the SOC is not enough to be repeatable and reliable enough to give me consistent enough data. So what I actually do is once it reaches steady-state, I reset the trip 2 computer and let it calculate it out until the end of the 16.4-or-so mile run (or just before getting off the offramp because the deceleration will ALWAYS put it into EV mode, which then, the little bit of offramp will bump up the fuel economy (semi-artifically).

 

Haha....I guess I should have written about the testing methodology. haha... whoops (no time).

 

And because I have such a long and progressive run-up, the time it takes to reach what I call the "full-at-speed" SOC is really, really short.

 

The other reason why this works even though it the statement is true: suppose you start off with a low(er) number because it's still recharging the battery to get it up to the "full-at-speed" SOC. It becomes a limit function, which means that the actual will be negligbly higher. (Not at the resolution that the trip computer can display anyways). So, suppose say...you start off at like sub-10 mpg. Once you're on the highway, if it ultimately will settle in around a steady-state of 40 mpg, your initial dy/dx will be a lot higher than as it approaches the steady-state solution. At that point, your d^2y/dx^2 is going to be a LOT smaller, to the point that while the statement is true, it doesn't matter.

 

That's why that the elevation and elevation changes was so important for me to note. IF there had been more changes in elevation, I wouldn't have been able to do that, because I wouldn't be able to quite hit a true steady state. And I mean, you can see pictures of the highway in my album, and you can see how it's just flat-flat-flat (almost as far as the eye can see).

 

And I'm not looking at the instantaneous fuel economy. (It's REALLY hard-as-in-nearly-impossible to tell on the instantaneous gauge whether you're a 37.3 or 34.3).

 

 

Is the utility of the regression supposed to lie in predicting fuel economy on that stretch of road at different speeds?

 

It doesn't seem to give very realistic values when I try it at slower or faster speeds.

 

Well, the key to it is that you have to be able to STAY at that speed. Trying to find a road that straight and flat and no traffic management stuff (stop lights, stop signs, etc.) where you can drive at lower speeds (safely, and without holding up traffic in the middle of the night) is quite difficult actually.

 

I thought about dropping down to 80 km/h (50 mph), but it's too slow for the flow of traffic for that highway, even if I were to do it in the middle of the night where there isn't much in the way of "traffic".

 

And while I can change it to some county road or something, the fundamental testing has changed because the route that I am taking has or is changing.

 

I WOULD test it faster, except that in Ontario, I think doing 30 over the limit is like $295 and 3 or 4 demerit points and the OPP had special semi-marked cruisers that were made the EXACT same shade as night (and speed radars are also illegal in Ontario).

 

So as much as I love science, it's not worth the risk. And there isn't a section of the highway in and around the Southeast Michigan area that's super flat like that that I can do a higher-speed run on; which again, by changing the road, I'm no longer consistent in my testing methodolgy which will instantly raise the question is the change of the fuel economy due to speed or due to the change in the test (which just creates more work trying to figure out the weight function coefficients between those two variables).

 

So, that rules that out.

 

Keep in mind that this is MY master equation because of the options that I have installed on the vehicle (which affects the mass of the vehicle) which ultimately, affects its fuel economy.

 

It is not meant to be a universal master equation or a family of equations or a collapse of a family of equations onto one that governs the speed vs. fuel economy for ALL FFHs.

 

But it does very clearly show the effect of speed vs. fuel economy.

 

And the other thing that I was thinking about more as well was "why didn't the electric motor kick in more at < 100 km/h?" and that might have something to do with how I normally drive the car and because it has "learned" that about me and so it's got the whole "dynamic calibration based on your driving style" thing going on as well. The only way that I would be able to probably force it to switch over to EV mode is if I slowed it down by like 1 or 2 km/h to force the switchover, and then speed back up to cruising, so that I would approach a geometric mean that's very close to if I had ran pure steady-state.

 

And I also didn't do the test for longer distances because I cannot guarantee that the highway remains that flat for the entire duration of it. And the time factor (cuz remember, I'm doing multiple runs. Doing (at least) 40 miles at 90 km/h would take me almost an hour JUST for that run. (And on a road THAT flat, you can see it settle in on a number with no visible fluctations actually like...LONG before my current run distance is complete.)

Edited August 9, 2013 by alpha754293

[acdii]

 

(It's REALLY hard-as-in-nearly-impossible to tell on the instantaneous gauge whether you're a 37.3 or 34.3)

 

Not with one of these. http://www.scangauge.com/products/scangaugeii/ Very accurate once calibrated. When I used it in that "other" car, it was within .1 MPG of actual. Supposedly the codes are out now to read SOC too, but haven't bothered to try it since I don't need to on the car I have now.

 

I might use it this winter once I put grill covers on just to keep an eye on engine temps. Right now its in the glove box of my Flex in case I need to tow something so I can keep track of engine and trans temps.

[mokee]

For what its worth...

 

http://fordfusionhybridforum.com/topic/6503-scangauge-ii-x-gauge-codes/

[alpha754293]

Not with one of these. http://www.scangauge.com/products/scangaugeii/ Very accurate once calibrated. When I used it in that "other" car, it was within .1 MPG of actual. Supposedly the codes are out now to read SOC too, but haven't bothered to try it since I don't need to on the car I have now.

 

I might use it this winter once I put grill covers on just to keep an eye on engine temps. Right now its in the glove box of my Flex in case I need to tow something so I can keep track of engine and trans temps.

 

None of the codes that have been posted pertain to actual fuel (unless you don't need to program those yourself).

 

*correction* By DEFAULT, its impossible to read the instantaneous fuel economy off the IP cluster. You have to rely on something else to read it.

 

And why would you care what the instantaneous is anyways? (never understood why people bothered with that). If you can't sustain it, it's meaningless. Your RAFE is probably the more important number.

 

If I were to be tracking instantaneous values, I'd be MUCH more interested in FF or FFpH.

 

(Not sure why you quoted the fuel economy numbers for 110 km/h and 120 km/h respectively...)

 

For what its worth...

 

http://fordfusionhybridforum.com/topic/6503-scangauge-ii-x-gauge-codes/

 

Like I said, the limit washes out what SOC you start from.

 

I'd only care about the instantaneous fuel economy if I was doing the data processing myself (real-time integration). And I think that you'll find that as your delta-t's get smaller, the more of an overkill doing that would be.

 

And considering that using just the trip computer, I'm already able to get an R^2 of 0.99785, I don't see the value in doing all of this "extra" for that 0.00214 (which, in theory, I can just choose a higher order polynomial to fix that, or perhaps use a different regression algorithm instead of LSE).

[hybridbear]

And why would you care what the instantaneous is anyways? (never understood why people bothered with that). If you can't sustain it, it's meaningless. Your RAFE is probably the more important number.

I agree! Alpha, with your connections can you please find out the time period used by the car to calculate the RAFE? I really like this statistic in the ET Mode but don't know the timetable behind it which limits its value sometimes. Thank you!

[milleron]

Alpha, perhaps you misunderstood my question. I'm asking if you did the regression analysis in order to be able to predict what mileage you might get at other speeds you could plug into your equation. I do understand that you cannot test it at all speeds.

More simply, why did you perform this statistical analysis?

Edited August 11, 2013 by milleron

[IraF]

169.8 miles Round Trip on I-64 Between Lexington KY and Clarksville IN with AC on @ 72Deg F while travelling aprroximately 68mph +/- 2-3 miles mph with approximately 55mph construction zone (5miles)

Actual Fill up 3.6 Gal

 

After Filling up and travelling approximately 4 miles in the City with Stop and Go traffic

Updated PCM and Gen2 v3.6.2 applied

 

 

 

 

 

Edited August 12, 2013 by IraF

[alpha754293]

169.8 miles Round Trip on I-64 Between Lexington KY and Clarksville IN with AC on @ 72Deg F while travelling aprroximately 68mph +/- 2-3 miles mph with approximately 55mph construction zone (5miles)

Actual Fill up 3.6 Gal

 

After Filling up and travelling approximately 4 miles in the City with Stop and Go traffic

Updated PCM and Gen2 v3.6.2 applied

 

 

 

 

 

hwy.jpg

Hwy Lifetime Summary.jpg

city.jpg

 

NICE! I kinda can't wait until I get my PCM and MyFord Touch updated. I might end up being SLIGHTLY under the 10k mark by this weekend when I get all that stuff done, but it should be good.

 

I also kinda wished there was like a "charge depletion mode" where you can force it to allow you to go up to like 60% power (on the 'Empower' screen) in electric mode, so you purposely delay the ICE kicking in.

[IraF]

Might mention had I turned off AC while travelling RT between KY and IN I am sure I would have done pretty close to 50mpg, but I can say for sure that 4 miles in the city I did turn off AC and had the Driver side as well as the Passenger side windows down

Edited August 12, 2013 by IraF

[dalesky]

Not with one of these. http://www.scangauge.com/products/scangaugeii/ Very accurate once calibrated. When I used it in that "other" car, it was within .1 MPG of actual. Supposedly the codes are out now to read SOC too, but haven't bothered to try it since I don't need to on the car I have now.

 

I might use it this winter once I put grill covers on just to keep an eye on engine temps. Right now its in the glove box of my Flex in case I need to tow something so I can keep track of engine and trans temps.

A friend of mine has one of these hooked up in his Honda truck. It gives an amazing amount of info, not the least of which is the awful fuel economy his truck gets!

[alpha754293]

Might mention had I turned off AC while travelling RT between KY and IN I am sure I would have done pretty close to 50mpg, but I can say for sure that 4 miles in the city I did turn off AC and had the Driver side as well as the Passenger side windows down

 

Were there a lot of elevation changes?

 

I've noticed that on flat land (like COMPLETELY flat land), it might take me a little while for me to "re-train" my FFH to learn to switch over to EV mode at the higher speeds (because my last test, it wouldn't do it like I thought that it would when I was driving at or < 100 km/h (62 mph).

 

A friend of mine has one of these hooked up in his Honda truck. It gives an amazing amount of info, not the least of which is the awful fuel economy his truck gets!

 

Well, the thing that I am "worried" about is whether the CAN bus signals for the MY2013 Fusion Hybrid was reversed engineered or whether it was unofficially released (cuz according to the guys on the OpenXCPlatform, it's still protected/proprietary/secret) so how it got out is perhaps a little alarming.

[IraF]

Up and down hills, but nothing which was steep and one might characterize as what you would expect. (Minor)

[alpha754293]

Up and down hills, but nothing which was steep and one might characterize as what you would expect. (Minor)

 

The reason why I'm asking is because I've observed that on the downhills, even little ones, it would be enough to make the EV mode kick in. But as a said, on perfectly flat land, with the way that I normally drive on my normal commute to/from work; it doesn't (kick over to EV mode). I have to lift off the throttle (completely) for it to do that.

[IraF]

My observations, going downhill is where EV mode kicks in (cruise on) and on flat land with cruise on it goes into EV mode except when the battery need recharging.

[MXGOLF]

I just finished a 861 mile trip at 68 - 70 MPH and got 35.4 mpg. This is with the new upgraded software. there 47 mpg advertised for my car is BS. I still think and know now that I got a lemon. :(

[corncobs]

C'MON FORD

listen to the people that have underachieving FFHs figure out what's wrong with them and make an awesome product even better by fixing these issues.

[alpha754293]

My observations, going downhill is where EV mode kicks in (cruise on) and on flat land with cruise on it goes into EV mode except when the battery need recharging.

 

It might take some time for my car to "re-learn/re-train" when to switch over to the EV mode and when not to after I get the PCM reprogrammed, hopefully this Saturday.

 

But yes, going downhill, even a very tiny one, with the cruise on for the whole drive, should help you switch over to the EV mode. Mine won't do it right now because the top speed is still limited to 100 km/h, and I set the cruise at 110 km/h.

 

I just finished a 861 mile trip at 68 - 70 MPH and got 35.4 mpg. This is with the new upgraded software. there 47 mpg advertised for my car is BS. I still think and know now that I got a lemon. :(

 

How many miles do you have on your car so far? It does take a while for the car to break in, I've found. Especially in the first 5000 miles. It's been easier for me to hit 40+ mpg, while holding at 70 km/h (44 mph) and from the moment that I get off the highway, to when I get into the office is about 2.1 miles, and if I can hit all the green lights, I can do it on a fully charged battery (charged via regen coming off the highway offramp), A/C off.

 

C'MON FORD

listen to the people that have underachieving FFHs figure out what's wrong with them and make an awesome product even better by fixing these issues.

 

Well, the fundamental fact is that the US EPA tests were originally developed for gas engines only, not so much EVs and hybrids. So, we test to what is required of us by federal law.

 

Unfortuantely, the side effect of that is the discrepancy between what's called "real world" fuel economy and the EPA rating/sticker. But as I have also said elsewhere, my answer to that is "if you had to design a test that captures all of the ways that people drive, in all of the places that people drive, in all of the environmental conditions that people drive, ranging from Alaska to Arizona, from Florida (lowest elevation) to Colorado, from Manhattan to middle of Iowa - and the test or tests have to answer the question "what is my fuel economy?" and it has to be an apples-to-apples comparison across the board, across all of those variables (and the total test or tests can't take too much time to run) - what would you suggest the test oughta look like/be?"

 

Of all the people that I've posited that question too, no one has been able to come up with an answer that don't have the same kind or range of problems that the current testing standard has.

 

It's one thing to point out the problems. It's another thing if people that point out the problem, can also point out the solution that'll fix everything (since that will actually address their own concerns).

 

And the test has to be able to cover cars ranging from the Scion iQ all the way to the Bugatti Veyron, and it has to cover all of the fuel/energy sources/types as well as whatever may come in the future.

 

Not so easy, huh?

 

"If it were easy, everybody would be doing it."

[alpha754293]

The highest fuel economy I've seen on fuelly is like 149.6 mpg or something ridiculous like that. Since I can't verify that, the next highest average that I've seen is like ~70 mpg. So there are people that CAN do better than the EPA sticker value. And there's a whole 'nother group of people that have been getting worse.

 

And while I rarely hit 47 mpg, but my personal goal was to be able to hit at least 40 mpg, which from the last picture that I posted up on my gallery here, shows that I can do that, with very little EV miles, still cruising at 110 km/h, over 114 miles (which WASN'T easy to accomplish) over such long distances.

[corncobs]

The highest fuel economy I've seen on fuelly is like 149.6 mpg or something ridiculous like that. Since I can't verify that, the next highest average that I've seen is like ~70 mpg. So there are people that CAN do better than the EPA sticker value. And there's a whole 'nother group of people that have been getting worse.

 

And while I rarely hit 47 mpg, but my personal goal was to be able to hit at least 40 mpg, which from the last picture that I posted up on my gallery here, shows that I can do that, with very little EV miles, still cruising at 110 km/h, over 114 miles (which WASN'T easy to accomplish) over such long distances.

But 40 MPG is still not EPA and I guess people would be fine with that based on real world driving but mxgolf's FFH can barely makes 35 MPG. A similar issue we saw with acdii's BD which he has now proven multiple times wasn't caused by his hybrid driving skills because those are excellent.

[IraF]

The Ford Fusion Hybrid 2013 SE is my second Hybrid that I own after trading in the Civic Honda Hybrid 2007 and at that time I believe there were only 2 hybrids on the road (Prius & Honda). Correct me if I am wrong.

When I owned the car I was extremely conservative in driving and I would commute about 60 miles round trip on very flat land and was able to obtain 50-60mpg. Of course windows open and driving 55mph. Honda's claim that year was 49/51 mpg. Many people complained that they were not achieving those numbers and they were later sued and were supposed to produce a video which you could watch on line to achieve those numbers or close to them.

The Honda at that time was using Nickel Hydride batteries and the engine was linked to the Hybrid battery and not like the Ford Fusion Hybrid today where you can use EV without causing the ICE to start up. Honda did claim under certain circumstances you could separate the engine from the Hybrid Battery, but I never observed that.

So I became interested in Hypermiling at the same time and started applying some techniques that I had read about and still use them today with the Fusion. In fact hypermiling can also be used with non hybrids also.

If possible use brake lightly when coming to a red light, going down hill put car in neutral and from my observations the mpg increased rapidly as well as inflate the tires to about 45psi. My first set of tires were replaced at 48,000 miles and I considered that normal. Never had the nerve for getting behind a 18 wheeler and letting him pull me. No cruise when going up hill.

I'm not telling you new revelations and of course we know in the winter the mpg will decrease due to a richer fuel mixture and this I have noticed with the fusion also.

A definition of a hill to me is and this is for the ones that have travelled between Jellico TN and Knoxville TN on I-75 you know what I am talking about or the Hill once you leave the visitor's center from the TN side of the Smokie's and head towards Cherokee NC or vice versa, those are hills.

My Hybrid battery failed me in February 2013 however Honda's warranty was 8 years/80,000 miles and since I had about 72,000 they replaced the battery with a refurbished battery(Frame was old, but the cells were new) and No charge.

My daughter has a 2009 Civic Hybrid and I believe with a Nickel Hydride battery and hers went bad also and no charge to her. Then I decided to trade.

The spec's with Ford Fusion Hybrid SE and comparing to a Prius V are like night and day to me and the Fusion was my only choice.

BTW my brakes on the Honda at 72,000 were 40% in the front and 50% in the rear.

 

 

This is from a Civic Honda Hybrid owner 2007

 

have noticed other people complaining about the mileage being poor, which I would attribute more to driving habits than capability of the vehicle. Where I live in Maryland is very hilly and I still manage over 49 mpg, sometimes over 50. It can be done, one just needs to understand that quick acceleration, trying to maintain 65-70 uphill, idling, warming up the car, etc, will cause mpg to go down (especially idling/warming up the car as you are registering 0 mpg the entire time). The car handles excellent around corners and in all kinds of weather and is easy on the tires (my original set lasted over 74,000 miles). While it doesn't specifically say it in the manual, I have found better performance is achieved when using premium gas. That really isn't a big issue because you're typically looking at a difference of $1-$2 on a fill up when compared to regular since the gas tank isn't big (I typically fill up around 9.2-9.5 gallons).

 

I am not here to criticize anybody's driving habits, but just to let you know the complaints are with other auto manufacturer's as well.

So can anybody explain how some do well and others don't?

[alpha754293]

But 40 MPG is still not EPA and I guess people would be fine with that based on real world driving but mxgolf's FFH can barely makes 35 MPG. A similar issue we saw with acdii's BD which he has now proven multiple times wasn't caused by his hybrid driving skills because those are excellent.

 

Well, it depends on the details about his vehicle. Has he broken it in yet (for the first 1000 miles per the instructions in the owner's manual)? How many miles does he have on the clock so far?

 

What are the options that he's put on his car (it's weight, and weight is bad for fuel economy).

 

For example, I think that my moonroof is like some 14 kg or something like that. And I think that for each of my 18" wheels, they're like an additional 5 kg at each corner. And compared to the non-hybrid Fusion, the battery is another what - like 200 kg? something like that. (I forget). My point is that it all adds up.

 

And while some people might say "what's 5 kg?" (where it seems like it's not a lot), but if you have 5 kg here, 6 kg there, pretty soon, you're 200 kg over and you're wondering "what the heck?".

 

It's all physics. (yayyyy physics).

 

I know that when I first got my Titanium Hybrid, I was barely getting 35 mpg as well. But as the miles started accumulating, the fuel economy got better. Like it used to be nearly impossible or seemingly impossible for me to hit or come close to 40 mpg on my commute to work (cruising at 110 km/h, there's one section where there's a big "hill" that it climbs, but then it comes back down off of it as well). Now, I've been able to hit it a few times or get a heck of a lot closer to it.

 

If you want the hybrid purely for fuel economy reasons (you don't care about looks or options), get the most basic SE hybrid you can get. One of the other guys at work says that he's been able to routinely get over 50 mpg on his commute to work, driving faster than I do, and further too.