1. Attachments are working again! Check out this thread for more details and to report any other bugs.

Accounting for computer discrepency in MPG Tests at various speeds

Discussion in 'Gen 3 Prius Fuel Economy' started by kgall, Jan 20, 2010.

  1. kgall

    kgall Active Member

    Joined:
    May 3, 2009
    984
    152
    2
    Location:
    Olympic Peninsula, WA
    Vehicle:
    Other Hybrid
    Model:
    N/A
    I'm hoping to test MPG at various speeds--setting cruise control at X and going for some way and then back, for example.
    The tests will be so short that it wouldn't make sense to refuel after each one--even if I did, the random error of refilling would be very large relative to gas used.
    So, I plan to use the computer-indicated MPG for each test, and adjust for the systematic high bias.


    I have seen some charts (Hobbit's maybe????? Other regulars here????) that treat the error as linear with speed--i.e., the error in MPG is about twice as big at 80 as 40--another way of putting it is that percentage error is constant.
    Given the method I plan to use, I will not be able to test whether this is in fact true.

    My question is:
    Is this based on empirical evidence, on some other information about how the MPG calculator works, or is it an educated guess?

    The alternative hypothesis I might have is that the MPG error remains constant.

    Any comments welcome.
     
    #1 kgall, Jan 20, 2010
  2. Indyking

    Indyking Happy Hyundai owner...

    Joined:
    May 28, 2009
    1,280
    90
    0
    Location:
    I don't know... Indy, Chicago, Madison (WI)... it
    Vehicle:
    Other Non-Hybrid
    Model:
    N/A
    The MPG error is not constant at all. Attached is a table with the variation from indicated vs. calculated after 28 HWY trips at average speeds varying from 54.1 to 64.7 MPH (each trip = 360 miles).

    Lowest variation is highlighted in green and highest in red.
     

    Attached Files:

    #2 Indyking, Jan 20, 2010
    1 person likes this.
  3. bwilson4web

    bwilson4web BMW i3 and Model 3

    Joined:
    Nov 25, 2005
    27,398
    15,524
    0
    Location:
    Huntsville AL
    Vehicle:
    2018 Tesla Model 3
    Model:
    Prime Plus
    The older NHW11 and NHW20s had a bladder in the fuel tank so we could never really trust the amount of fuel added. In my case, I just used the cumulative gas and calibrated trip meter distances to handle my actual MPG.

    Even though the fuel tank is fixed, I still notice the ability to squeek in more fuel. So mostly I don't worry about the computer display error. We did a statistical analysis last summer and it looks to be ~5% ... over the long haul. But it isn't something I lose sleep over.

    GOOD LUCK!
    Bob Wilson
     
    #3 bwilson4web, Jan 20, 2010
    1 person likes this.
  4. kgall

    kgall Active Member

    Joined:
    May 3, 2009
    984
    152
    2
    Location:
    Olympic Peninsula, WA
    Vehicle:
    Other Hybrid
    Model:
    N/A
    Bwilson,
    A while ago, you posted this chart:

    Model: III
    Package: #1
    Thanks: 210
    Thanked 532 Times in 315 Posts


    Friends: 20
    [​IMG] Re: Long Trip MPG
    Quote:
    Originally Posted by pakitt [​IMG]
    The Prius 2010 has been tested in Italy by "Quattroruote" car magazine . . . they use their own, pretty high tech and highly accurate, instrumentation):

    90km/h (56mph) 26,8km/L (63.04 mpg)
    100km/h (62mph) 23,4km/L (55.04 mpg)
    130km/h (80mph) 15,4km/L (36.22 mpg)
    (this is the italian highway speed limit and the advised speed limit for germany)
    150km/h (93mph) 11,7km/L (27.52 mpg)
    . . .

    [​IMG]

    Is this from your own experience or from Quattroruote?
    If from yours, Could you explain the two lines, and how you got them? It appears that the difference you indicate between actual and indicated mpgs is roughly linear with regard to actual MPGs. True?
    How did you get the difference at various speeds, as I doubt you were able to burn a whole tank of gas at constant speed at a track?
    Thanks a lot!
     
    #4 kgall, Jan 21, 2010
  5. bwilson4web

    bwilson4web BMW i3 and Model 3

    Joined:
    Nov 25, 2005
    27,398
    15,524
    0
    Location:
    Huntsville AL
    Vehicle:
    2018 Tesla Model 3
    Model:
    Prime Plus
    I recorded most of the data on the drive back home. I've added a few other points from Ken@Japan, a P&G test and an early morning, 80 mph test.
    The upper line is the indicated mileage. The lower, thinner line is the corrected value.
    It is my assumption. To really tell, I would need a 5th wheel for exactly accurate velocity and injector instrumentation.
    The protocol was:

    • use either I-65 or I-565 or access road along I-565 for the runs. I-65 is posted 75 mph, somewhat loosely enforced. I-575 is posted 70 mph and the access road is posted at 55 mph and sparse traffic.
    • accelerate to target speed and trim cruise control for target
    • identify landmark for return trip in opposite direction
    • reset trip meter
    • run 10 miles but throw out run if traffic requires a speed change
    • approaching exit, record miles, MPG, and average speed (had portable GPS to verify speed, was 1 mph off)
    • park long enough to make a written record
    • get back on divided highway repeat same protocol
    • combine opposite direction runs for each datapoint
    I've found that 10 miles is more than enough to get useful data. Shorter distances begin to have more variability. For the slow speeds, I used road on Redstone Arsenal on Saturday. Again, 10 mile runs or at slower speeds, a circular access road posted at 25 mph around a parking lot, ~1.1 miles per loop.

    The car was 'stock' from the dealer including their tank of Shell 89, E10.

    Bob Wilson
     
    #5 bwilson4web, Jan 21, 2010
    1 person likes this.
  6. ken1784

    ken1784 SuperMID designer

    Joined:
    Dec 18, 2003
    2,940
    1,365
    67
    Location:
    Yokohama, JAPAN
    Vehicle:
    2010 Prius
    Model:
    N/A
    I have seen very consistent discrepancy, +6.9%, plus/minus within 1% regardless speed (city or highway driving).
    I bet there is no relations between the discrepancy and speed.

    Ken@Japan
     
    #6 ken1784, Jan 22, 2010
    1 person likes this.
(You must log in or sign up to post here.)