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

Roadster Charge Time Predictor

Discussion in 'Roadster' started by tomsax, Dec 10, 2013.

  1. tomsax

    tomsax Member

    Joined:
    Jun 19, 2008
    Messages:
    867
    Location:
    Sammamish, WA
    Have you ever wanted to get an idea of how long a charge will take? Perhaps you've used the charge rate table from my ancient blog, then added 20% because you're charging at a 208V station instead of 240V, and then added in some extra time because you know the charge rate will taper down as it gets closer to fully charged. Been there, done that.

    I've wanted to improve that experience, but it took a bunch of time just to do the very simple analysis for that blog and I needed a lot more data. Earlier this year, MarkWJ gifted me with anonymized data collected from over 7000 charging sessions via OVMS from 126 vehicles. That data allowed me to build a much better charging model for the Roadster. The data also shows how much variation there is among charge sessions due to external factors, so the charge time prediction isn't going to be perfect but it seems to be working pretty well. How I developed the charging model and its limitations are explained here: OVMS and the Tesla Roadster Charge Time Predictor

    The charge time predictor is now built into the latest firmware version for both OVMS and the Tesla Tatter. OVMS now has a new feature (only available in V2 hardware) that lets you set a charge end time, so your car will finish charging close to when you're ready to leave (within the limits of the accuracy of the charge time predictor). It's also better for the grid if EV owners set a change end time instead of a charge start time to avoid the spike from every EV starting to charge when the lowest time-of-use rate kicks in.

    For Roadster owners who don't have one of the telematics gizmos, I highly recommend you get one, but I've also made the charge time predictor available via a web page:

    Tesla Roadster Charge Time Predictor

    You can fill in custom values (charge mode, battery pack capacity, preferred units, typical ambient temperature) and the page will give you the time estimate and also generate a link you can bookmark to load the page later with your values already filled in.

    The algorithm seems to work pretty well for level 2 charge rates in moderate temperatures. If you're doing 120V charging, especially at 12A and lower, the charge rate is highly variable and more difficult to predict. The predictor doesn't consider starting battery pack temperature, so it doesn't take into account situations where the car is cold enough that the pack needs to be heated before it can charge (I need more data to model that). I also expect it won't do a great job in really hot weather because the values from the ambient temperature sensor on the vehicle are often quite different from the actual air temperature. It also doesn't do anything to model the difference between charging in an open air carport and a small, tightly-enclosed garage. Still, it's better than the old way of estimating charge times.

    Please give it a whirl. Let me know how it works. If you have a case where it's consistently wrong, send me your data, preferably with the corresponding OVMS server logs.
     
  2. Doug_G

    Doug_G Lead Moderator

    Joined:
    Apr 2, 2010
    Messages:
    15,852
    Location:
    Ottawa, Canada
    Very cool, Tom! Give the man a little data, and he pops out all kinds of useful info.
     
  3. markwj

    markwj Moderator, Asia Pacific

    Joined:
    Apr 10, 2011
    Messages:
    3,665
    Location:
    Hong Kong
    If you check the internal diagnostics screens in a roadster, you'll see that a Tesla actually has a value for charge time remaining, but it is always -1. Most impressive that Tom managed to get this done, when Tesla couldn't / wouldn't.

    For most use cases, +/- 30 minutes seems plenty accurate enough, and Tom's algorithm easily meets that goal.

    Well done, Tom, and glad the data was useful to you. It is going to be really interesting, over the coming years, how such publicly released anonymised data can help the EV community, and OVMS will certainly do its part to help. In fact, we've got an announcement coming up real soon regarding this, that I'm hoping will produce some truly useful data.
     
  4. sethr

    sethr CPO Roadster #1089

    Joined:
    Mar 24, 2013
    Messages:
    152
    Location:
    Fairfax, VA
    This is really great. One small glitch in the web page time predictor: after entering a different ambient temperature, it still is showing 80 degrees.
     
  5. Pantera Dude

    Pantera Dude Member

    Joined:
    Mar 18, 2012
    Messages:
    468
    Location:
    Long Beach, California
    Near the bottom of the page there is a link for "Charge Time Predictor Custom Preset Values." It took me awhile to find it.:biggrin:

    - - - Updated - - -

    Outstanding job Tom! Very cool indeed!
     
  6. tomsax

    tomsax Member

    Joined:
    Jun 19, 2008
    Messages:
    867
    Location:
    Sammamish, WA
    After changing a value, you have to tab or otherwise select another field, or hit the Update button. If I do that, I see the time estimate, the summary text, and the link all change to reflect the temperature change. If you're having a different experience, let me know what browser and version you're using.
     
  7. ibcs

    ibcs Member

    Joined:
    Jan 31, 2010
    Messages:
    720
    Location:
    Ohio
    Thanks Doug - Love all the new stuff for the roadster. New CPO owners are just lucky to have all the experienced owners.
     
  8. hcsharp

    hcsharp Active Member

    Joined:
    Jun 7, 2011
    Messages:
    2,537
    Location:
    Vermont
    Having the same problem with Firefox 25.0.1. Many thanks for doing this Tom.
     
  9. tomsax

    tomsax Member

    Joined:
    Jun 19, 2008
    Messages:
    867
    Location:
    Sammamish, WA
    I found a problem with the Update button, now called Calculate instead. Let me know if you're still seeing a problem.
     

Share This Page