TMC is an independent, primarily volunteer organization that relies on ad revenue to cover its operating costs. Please consider whitelisting TMC on your ad blocker or making a Paypal contribution here: paypal.me/SupportTMC

Improving EV TripPlanner with our data

Discussion in 'Model S: Battery & Charging' started by Boatguy, May 18, 2016.

  1. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    One of the critical assumptions in EV TripPlanner is the Wh/RM of each vehicle; it is the foundation upon which all the other adjustments are made to determine power consumption for a trip. After running some scenarios, most of the models use 300 Wh/RM, which is also what Tesla tech support told me they use. But in looking at my own MS 90D w/19", I've concluded it uses 290 Wh/RM.

    I arrived at this by evaluating this equation half a dozen times and always getting 288-290 Wh/RM:

    30mi average Wh/mi / (Rated range / Projected range) = Wh / RM.

    Of course the calculations can also be done in km and Wh/km and then converted to miles.

    Applying the 290 Wh/RM to my 100% range of 292 yields a usable battery of 84.7 kWh which also seems about right (i.e., ~ 5 kWh battery reserve / buffer). By comparison my BMW i3 has a 3.2kWh reserve/buffer based on BMW's published numbers.

    The values are all displayed as integers so we're going to get a little more variation in the Wh/RM than if we had one decimal of accuracy, but I suspect they are going to cluster pretty tightly.

    I'm suggesting a three step process:

    1) Some of the more critical readers need to verify that the above formula will yield the assumed Wh/RM. If that flunks, then game over.

    2) If the formula is correct, we need a few owners of each model/wheel combination to do the math and determine the assumed Wh/RM for their car.

    3) We compile the results and send it to Ben Hannel so he can update EVTP with the corrected numbers. Then we'll all have more accurate projects for our road trips.

    Let's try step 1, does this make any sense?
     
    • Informative x 1
  2. TexasEV

    TexasEV Active Member

    Joined:
    Jun 5, 2013
    Messages:
    3,770
    Location:
    Austin, TX
    If Tesla told you 300 Wh/RM they were rounding off, as that's a reasonable approximation. It's a little different for each model. Those calculations have already been done, probably 2 or 3 years ago before the 90 came out. They're in the site here somewhere.
     
  3. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    The point is that our cars have different assumed Wh/RM and we can improve the accuracy of EVTP if we supply Mr. Hannel better information about the RM calculation in the different model/wheel variations.

    Until Tesla develops a decent trip planner, EVTP is the only game in town.

    So TexasEV, you've got a 60. What is your 30 mi avg Wh/mi / (rated range / projected range) = ??

    It's a five minute exercise to look at the screen while it's parked in your garage.
     
  4. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    You're basically introducing error from your personal driving area/style into average energy consumption. You should instead try and find out what YOUR Tesla thinks is Rated Wh/mi. There are several ways to arrive at that. One would be to drive in such a way as your actual consumption lines up with the rated line in the energy analyzer, and use that value.

    That said, it would be less labor/more flexible for evtripplanner to allow the user to set both rated Wh/mi and pack capacity, with the drop down for defaults for people that don't know.
     
    • Like x 3
    • Helpful x 1
  5. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    Can you please elaborate on that? My understanding was that Rated miles is entirely independent of my driving style.

    If EVTP proceeded as you suggest, what would you enter for a P90D's Wh/Rm and pack capacity?

    I believe by using Projected miles and my average consumption we're taking my driving out of the equation arrive at the Wh/mi used to calculate Rated miles on all 90D.

    AWDtsla - would you mind humoring me? Post the numbers for your P90D, I'll do the math.
     
  6. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    It is, but you're arrive at 290 somehow, I believe tesla is right about 300. 70D is 290 I know that. I would at least try to procedure I described above to see what energy analyser claims. The other way to do it is the use "energy added" vs rated miles added. There could be some error if you haven't done a deep discharge lately, I would at least make sure your measurement charge is within a narrower range than previous discharge/charge.

    Trip meter lies, don't use that consumption number independently. You should be able to calculate it in 2-3 ways and arrive at the same answer. If anything use the number of miles lost in rated range time estimate Wh/mi instead of displayed consumption.

    I use ~83.5kWh and 312Wh/mi. I think the actual value is between 310 and 315. So Tesla claims 270 miles range now, but I think this is the usual exaggeration on numbers. I've never seen a 100% charge over 267 miles.
     
  7. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    OK, reset. Let me try this again.

    EV Trip Planner needs to know Tesla's assumed Wh/mi for each model/wheel combination when they compute Rated Miles. Then it can apply adjustments for speed, temp, altitude, etc. Tesla knows SOC and usable battery capacity, so they divide by some assumed Wh/mi to produce Rated Miles remaining as follows:

    Rated miles = SOC * Usable battery / Wh/mi.

    We should all be able to agree on that. How else could they compute Rated miles?

    We want to know the assumed Wh/mi so using the information we have, we have an equation with four variables and one unknown:

    Projected Miles = Rated Miles / (Avg Wh/Mi / Rated Wh/Mi)

    If your actual driving average Wh/mi is the same as Tesla's assumption, then Projected = Rated. If your consumption is higher, then Projected goes down, if your consumption is lower then Projected goes up. Tesla gives us three of these numbers. Solving for the unknown we get:

    Rated Wh/mi = Avg Wh/mi / (Rated range / Projected range)

    Our personal consumption is irrelevant beyond the fact that it allows us to compute for the unknown. Vampire is irrelevant because Telsa doesn't consider it when computing consumption. Any number of different owners of a given model / wheel dimension with different driving habits will end up +- 1-2 the same Wh/mi when they do the math because the underlying assumption in that vehicle is the same (assuming the wheels are factory installed).

    One other owner has already given me his numbers for a 90D/19" and it's within 1-2 Wh/mi of what I see for my 90D. His number was 288, my 90D/19" is in the range of 288-290 based on half a dozen observations. That's not a coincidence, it's because Tesla's Rated Wh/mi number in our two cars is the same.

    For EV TripPlanner to do a decent job, it needs to know Tesla's Rated Wh/mi for each model/wheel combination. Per above, it's trivial to calculate, we just need some observations from owners.

    Even if you don't agree, it only takes a couple of minutes to jot down three numbers and report them here:

    Model/wheels
    Rated Miles
    Projected Miles
    Average Wh/mi

    Thank you for indulging me.
     
  8. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    Can you confirm what your energy analyzer reports as it's "Rated" line?
     
  9. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    This morning (screen shots below), charged to 90% last night, my car shows:

    264 RM
    342 Wh/mi
    222 Projected range.

    That's a 100% charge of 293, well within rounding error of the 294 that Tesla claims; lucky me.

    The works out to a Rated Wh/mi of 288, which is exactly what the math shows for another 90D owner who sent me his numbers last night.

    Since the 90D is 288, and the P90D has a rated range of 270, then 288 Wh / mi / (270/294) = 314 Wh/mi which what you estimate for the P90D. Would you mind cross checking it with the rated range, Avg Wh/mi and Projected range approach?

    The point of all this is that EVTP should be using 288 for the 90D/19" and 314 for the P90D/19". If other owners send in their three numbers, I can do the math and forward them to Mr. Hannel so he can update EVTP and we'll all have more accurate trip plans.

    IMG_1016.JPG IMG_1017.JPG
     
  10. donauker

    donauker Member

    Joined:
    Sep 5, 2006
    Messages:
    755
    Location:
    Pennsylvania
    What gives you the idea that this hasn't already been done by the writers of EV TripPlanner? Based on 10s of thousands of miles of experience with both my P85 and my Roadster they have done a excellent job of very accurately reporting results for 2 very different vehicles!

    For some reason I am having difficulty seeing what you are looking for that is different than what the developers of the site started collecting over 2 years ago. Have you seen the following post by the developers?

     
    • Informative x 1
  11. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    Don - I had an exchange with Mr. Hannel, but perhaps I misunderstood and he just wanted data for the 90D/P90D.

    Something is not quite right in his calculations. If you run the same route in a P85D and an 85D, he computes a different total energy, but then converts the total energy to RM using 300 Wh/mi for both cars. I don't understand how he's getting the total energy if he's assuming the same Wh/mi for both cars. I'll check back with him.

    Where was that post and were the replies posted in the thread, or were they to be sent in?
     
  12. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    Can you describe that process exactly?
     
  13. donauker

    donauker Member

    Joined:
    Sep 5, 2006
    Messages:
    755
    Location:
    Pennsylvania
    It would make sense to me that they are looking for data for the 90D/P90D as that these models are not yet in their drop down list.

    You can access that thread by clicking on the little red up arrow next to his name in the quote in my previous post. There were some responses in that thread but I know there were other areas that they collected the data as well. I know the Roadster was added at a later time.
     
  14. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    FYI calculation via using remote S while charging: 8.28 kWh, 26.5 mi added 8.28=26.5=312Wh/mi.
     
  15. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    Yes. Jot down these three numbers:

    Instrument panel: Rated miles
    Consumption display on CID: Average Wh/mi, Projected range

    Then plug them into this equation: Rated Wh/mi = Avg Wh/mi / (Rated range / Projected range)
     
  16. AWDtsla

    AWDtsla Active Member

    Joined:
    Mar 3, 2013
    Messages:
    2,133
    Location:
    NE
    >>> 404./(227./172)
    306.11453744493394

    >>> 373./(227./187)
    307.273127753304

    >>> 386./(227./180)
    306.07929515418505

    So it reads low for my car. That's confirmed with a 100% charge, the pack capacity would have to go under 83kWh for the math to work. Also, the CID is very prone to error, it can miss sampling intervals throwing everything off.
     
  17. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    This is a P90D and your 100% rated range is 227?

    At 306-307 the Wh/RM is lower than I would have expected, but obviously it's consistent. Thank you.
     
  18. ShotgunF15E

    ShotgunF15E Member

    Joined:
    Jan 30, 2016
    Messages:
    337
    Location:
    San Diego, CA
    I get 270 every time I do a 100% charge.
     
  19. Boatguy

    Boatguy Member

    Joined:
    Apr 3, 2014
    Messages:
    675
    Location:
    SF Bay Area
    What do you get for Wh/RM?
     
  20. brec

    brec Member

    Joined:
    Aug 3, 2015
    Messages:
    287
    Location:
    Reno, NV
    85D/19"
    178.76
    143.97
    356
     

Share This Page