Hey guys. I've been fiddling with FE on my Prius and stumbled onto something that may already be well known, but wanted your thoughts on.

Have you noticed a sweet spot in fuel efficiency at around 47 mph. What I found is that around 47 mph the FE is better than it was at 42, and better than it was at 50. Which leads me to believe that there is a localized maximum on FE between 42 and 50.

For what it's worth, with cruise control on at 47mph, I seemed to be getting over 60 mpg, and at times over 70 mpg.

 

Hey, Dan. That's an interesting finding that I haven't noticed, though I rarely use CC these days.

You might check out some of Hobbit's work that he has described and linked on CleanMPG. He seems convinced that as CC tries to maintain speed it becomes too aggressive for optimal fuel economy, with wide swings of RPM deep into inefficient ranges, low and high. (It is based on my experience, validated by his findings, that I usually avoid CC.) His most recent paper describes what he calls a "forbidden zone," where a certain combination of mid-range vehicle and engine speeds puts the HSD into an inefficient mode. Your results could be from this same phenomenon, where bumping up to 50 from 47 nudges the vehicle speed and RPM together just out of this zone. The MPH/RPM chart in his paper would seem to support this.

I'm guessing based on your findings and where you live that most of your CC driving is on level terrain. Am I correct? I would wager that your mileage would be noticeably worse with a CC-regulated 47 MPH on hills.

 

Hi Dan,

Excellent! Can you describe your methodology?

- distance
- level and/or slope
- temperature
- number of runs

I have an NHW11, previous model, and since picking it up in October 2006, I have been making a study of MPG vs. MPH:


What the above chart shows is the dip that happens at and around 42 mph. For both of our Prius, this is the transition speed between: (1) alternating gas and electric mode, and (2) gas must run mode. If you think 47 is great, you should try it at 38 or slower.

Over a flat, 2.5 mile, controlled section:


This chart is important because it shows what happens when you don't transit 42 mph. The test was very carefully crafted on a flat stretch without wind or elevation changes. This is pretty close to the limit of what the car can achieve at each speed.

But more recently, I've been working on a math model and gotten:


You'll notice there is very good agreement with the measured, 2.5 mile benchmark results. Since then, I've investigated 15 mph and achieved:



This agrees very nicely with what the model predicts. Best of all, I can hold the vehicle in this range as long as I want.

If you'd like some more details and metrics:
http://hiwaay.net/~bzwilson/prius/pri_test.html

BTW, Hobbit is an excellent source. He is pretty good about taking measurements and has done a lot of admirable work:
http://techno-fandom.org/~hobbit/#prius

What I like most is Hobbit's attention to detail and measurements. Sounds like you may have interest in that area and we sure could use the company. Another favorite hang-out is:

Prius_Technical_Stuff - a Yahoo group, http://groups.yahoo.com/

Bob Wilson

 

Uhh... Yikes... May be over my head here. May be an anomaly of the break-in, but I'll see what I can do. I noticed it a few weeks ago that my MPG was kinda low at 42 MPH. I figured that it should be high at 42 MPH since it would ensure I ran off the ICE and the wind resistance would be lower. I tried it at 50 and got about the same results. Over the next few weeks I tried 43,44,45,46,47,48,49. Seemed 47 was the best.

I'm afraid it wasn't very scientific. There is a 2 mile stretch, fairly level, with little wind. I simply watched the FCD and looked at what it bounced between.

I obviously need to find a good stretch without lights or stop signs. My main problem is that I don't have a Scangauge and I don't want to reset my tank average. I figure I'd need at least a 10 mile stretch to reduce error.

 

Not really, you're showing exactly the type of experimental attitude needed to move from FUD and superstition to knowledge and enlightenment. This is the way to go and you are in good company.

Actually this is pretty good. Use the resources readily at hand and build upon them. My 'benchmark' stretch is 2.5 miles long. Heck, my initial, high-speed benchmarks were on I-20 between Fort Worth Texas and Birmingham Alabama.

The instrumentation is nice to have but your multi-function display is perfectly adequate.

In my case, mileage is calculated using gas receipts and trip meter miles. I use the MFD for benchmarks and experiments. But whatever works, as long as there is some reasonable metric involved, is perfect.

Bob Wilson

 

Dan, any updates on your experiments?

 

Well I tried it this morning, but the engine was cold, so it was in the 50s.

One problem I do have is rounding error. Since I use the FCD to maintain my tank average, the only way I can measure fuel consumption is:

(Me - Ms)/(Ge - Gs)

Where:
Ms = FCD MileageStart
Me = FCD MileageEnd
Gs = GallonsConsumedStart
Ge = GallonsConsumedEnd

To solve for Ge and Gs I use the following:
Ge = Me/MPGe
Gs = Ms/MPGs

Where:
MPGs = FCD mpgStart
MPGe = FCD mpgEnd

Now enter rounding error. If MPG reads 53.5, it could actually be anywhere between 53.45 and 53.54 (both of which round to 53.5). Same with mileage. 474.3 could be anywhere between 474.25 and 474.34.

So if Ms and Me are both large numbers (as they are at the end of a tank) then the swing accounting for rounding over a 2 mile track can be huge. In this particular example I believe the swing comes out to anywhere between 39mpg and 81mpg based on a 53.5 mpg read.

So unfortunately, until I can find a sufficient track in the Houston area, I don't know If my measurements will be very helpful.

I am working on a new spreadsheet to give me the range that the values are in accounting for rounding, but I'll have to get some data first.

One thing that might help... does anyone know if the FCD does rounding or truncation? I assumed rounding, but it may truncate based on my trip -vs- FCD mi numbers.