Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Lifetime Average Wh/mi

This site may earn commission on affiliate links.
so here are my numbers
I am doing an experiment to see how low wh/m I can get.
it seems to be 230, no ac and very slow acceleration (try to keep it below 10kw ) around town driving
last charge was sun am, to 234 miles , today is Thursday, my A number represents a full electric bill cycle
230-.jpg
 
May numbers.

One thing I've noticed is that because of rounding errors in the reported wh/m numbers, there can be slight inaccuracies in my calculation of the monthly wh/m numbers for people who don't report these number separately.

For the monthly plots, I calculate wh/m_monthly = (tot_energy_to_date - tot_energy_as_of_last_update)/((tot_miles_to_date - tot_miles_as_of_last_update) where
tot_energy_to_date = avg_whm_to_date * miles_to_date, etc. This formula is exact, but slight errors creep in because avg_whm_to_date is
rounded. This can cause problems when the number of miles driven in the last month is much less than the total miles driven. For example, if you've
driven 20k miles, and the average wh/m is 330 and then in the next month you only drive 100 miles, say with wh/m of 300, then because of rounding, the car will still probably say your average wh/m is 330 even though it is now slightly less than that. Over time this should wash out, but any given month could be off a bit.

I'll try to be more careful about keeping track of monthly numbers for anyone who reports monthly numbers separately. (I do this already in many cases but haven't been totally systematic about it).

On the total whs coming out of the wall vs whs reported by the car question, phx182flyer keeps track of his whs from the wall on a daily basis. He's got a nice regression of wh_wall vs miles here:

Graphs: ElecPwr

He is basically seeing 365wh_wall/mile + 2 wh/day. This compares to his car which reports around 303 wh/m. I'd interpret this as charging losses of about 20%, plus 2wh vampire. This is in the same ball park as what mknox is seeing and Cottonwood's calculation.
 

Attachments

  • teslaWHM201405.pdf
    426.9 KB · Views: 135
May numbers.

One thing I've noticed is that because of rounding errors in the reported wh/m numbers, there can be slight inaccuracies in my calculation of the monthly wh/m numbers for people who don't report these number separately.

For the monthly plots, I calculate wh/m_monthly = (tot_energy_to_date - tot_energy_as_of_last_update)/((tot_miles_to_date - tot_miles_as_of_last_update) where
tot_energy_to_date = avg_whm_to_date * miles_to_date, etc. This formula is exact, but slight errors creep in because avg_whm_to_date is
rounded. This can cause problems when the number of miles driven in the last month is much less than the total miles driven. For example, if you've
driven 20k miles, and the average wh/m is 330 and then in the next month you only drive 100 miles, say with wh/m of 300, then because of rounding, the car will still probably say your average wh/m is 330 even though it is now slightly less than that. Over time this should wash out, but any given month could be off a bit.

I'll try to be more careful about keeping track of monthly numbers for anyone who reports monthly numbers separately. (I do this already in many cases but haven't been totally systematic about it).

On the total whs coming out of the wall vs whs reported by the car question, phx182flyer keeps track of his whs from the wall on a daily basis. He's got a nice regression of wh_wall vs miles here:

Graphs: ElecPwr

He is basically seeing 365wh_wall/mile + 2 wh/day. This compares to his car which reports around 303 wh/m. I'd interpret this as charging losses of about 20%, plus 2wh vampire. This is in the same ball park as what mknox is seeing and Cottonwood's calculation.

What an impressive write up! That may well be publishable somewhere. On one of the big EV blogs at a minimum. Thanks for all you do to track this data, it is greatly appreciated.
 
Thanks Efusco.

It occurred to me that the easiest way to avoid the rounding errors I noted above is simply to keep track of energy use directly. So in future updates if people post their energy totals (some people already do, and it's obviously also in the screen shot), I'll record that and it to calculate monthly numbers. I'll also keep track of the monthly numbers separately.
 
May numbers.

One thing I've noticed is that because of rounding errors in the reported wh/m numbers, there can be slight inaccuracies in my calculation of the monthly wh/m numbers for people who don't report these number separately.

For the monthly plots, I calculate wh/m_monthly = (tot_energy_to_date - tot_energy_as_of_last_update)/((tot_miles_to_date - tot_miles_as_of_last_update) where
tot_energy_to_date = avg_whm_to_date * miles_to_date, etc. This formula is exact, but slight errors creep in because avg_whm_to_date is
rounded. This can cause problems when the number of miles driven in the last month is much less than the total miles driven. For example, if you've
driven 20k miles, and the average wh/m is 330 and then in the next month you only drive 100 miles, say with wh/m of 300, then because of rounding, the car will still probably say your average wh/m is 330 even though it is now slightly less than that. Over time this should wash out, but any given month could be off a bit.

I'll try to be more careful about keeping track of monthly numbers for anyone who reports monthly numbers separately. (I do this already in many cases but haven't been totally systematic about it).

On the total whs coming out of the wall vs whs reported by the car question, phx182flyer keeps track of his whs from the wall on a daily basis. He's got a nice regression of wh_wall vs miles here:

Graphs: ElecPwr

He is basically seeing 365wh_wall/mile + 2 wh/day. This compares to his car which reports around 303 wh/m. I'd interpret this as charging losses of about 20%, plus 2wh vampire. This is in the same ball park as what mknox is seeing and Cottonwood's calculation.

Awesome! thanks again Jeff for putting the numbers together, and for pointing to the wall to wheel numbers. It makes me wonder how this compares to other plug in's? I'll have to try and hit up a leaf forum to see what kind of numbers they are getting for comparison sake.
 
Just following back with additional info on charging efficiency for other manufacturers.

http://www.mynissanleaf.com/viewtopic.php?f=31&t=8583
http://www.mynissanleaf.com/viewtopic.php?f=8&t=10307 <- links to an NREL investigation of the Leaf and includes the following graphic which seam relevant.
leafsystemefficiency.png


I realize this is slightly off topic here, so if the mods feel free to move or anyone can let me know where to cross post. BMW i3 info seems like the other logical comparison, I'll see what I can find.
 
Just following back with additional info on charging efficiency for other manufacturers.

http://www.mynissanleaf.com/viewtopic.php?f=31&t=8583
http://www.mynissanleaf.com/viewtopic.php?f=8&t=10307 <- links to an NREL investigation of the Leaf and includes the following graphic which seam relevant.
leafsystemefficiency.png


I realize this is slightly off topic here, so if the mods feel free to move or anyone can let me know where to cross post. BMW i3 info seems like the other logical comparison, I'll see what I can find.

That's interesting. Just as another data point, phx182flyer was kind enough to let me look at his charging data. As expected, I get the basically the same regression coefficients that he got when I repeated his single variable regression, total_wh_wall ~ a * miles +b with the same very high adjusted R^2 (93 ish). However I get a slightly better fit using with the regression total_wh_wall ~ a * miles + b * days, where days is the number of days covered by a given reading. Mostly days = 1 in his data, but there are some multi-day periods (if every observation covered one day, the two regressions would obviously be equivalent). The miles + days regression has the coefficients

total_wh_wall ~ .344 * miles + 3.4 * days

with the interpretation that .344 kwh = energy used to drive 1 mile (reported by car) + charging losses while 3.4 kwh = daily vampire draw. phx182flyers car reported wh/m was I believe around 304 in this period, so that indicates charging losses of (344-304)/344 = 12%. This is pretty close to the 86.69% charging +ESVE efficiency of the Leaf in the diagram you posted.
 

Attachments

  • rt2-2.jpeg
    rt2-2.jpeg
    85.7 KB · Views: 303
That's interesting. Just as another data point, phx182flyer was kind enough to let me look at his charging data. As expected, I get the basically the same regression coefficients that he got when I repeated his single variable regression, total_wh_wall ~ a * miles +b with the same very high adjusted R^2 (93 ish). However I get a slightly better fit using with the regression total_wh_wall ~ a * miles + b * days, where days is the number of days covered by a given reading. Mostly days = 1 in his data, but there are some multi-day periods (if every observation covered one day, the two regressions would obviously be equivalent). The miles + days regression has the coefficients

total_wh_wall ~ .344 * miles + 3.4 * days

with the interpretation that .344 kwh = energy used to drive 1 mile (reported by car) + charging losses while 3.4 kwh = daily vampire draw. phx182flyers car reported wh/m was I believe around 304 in this period, so that indicates charging losses of (344-304)/344 = 12%. This is pretty close to the 86.69% charging +ESVE efficiency of the Leaf in the diagram you posted.

so where were we going wrong in the previous posts when we all seemed to agree that the average losses were in the 16-17% range? or your previous post where you were seeing 20% charging losses + 2kwh of vampire? Just small sample size? variance in vampire numbers? variance in car to "wheel" efficiency? that last one we should at least be able to give some guidance on based on the info collected in this thread.