Trying to Understand Efficiency

[AlanSubie4Life]

@Nate977p

Did a quick check on my way home tonight (it is uphill so consumption is a little high). I stopped at the store, so it is in two segments. Note I did not use HVAC for this one so that is something to consider if your other numbers don’t align well (but I don’t think it matters). Another thing is that the car had just been charged, so it definitely was not cold-soaked, though regen was limited due to the high SoC of 95% starting point. I also launched the car once to 60mph, just to get a brief very high current event in the discharge profile:

Start 293 rmi
End 280 rmi

Segment 1: 4.7mi, 279Wh/mi. End 288 rmi
Segment 2: 4.8mi, 393Wh/mi. End 280 rmi

Summary :

Rated miles:

Consumed 13 rmi, 242Wh/mi = 3.15kWh

vs. trip meter:

4.7*279 + 4.8*393 = 3.20kWh. (Really there are only two significant figures here, so 3.2kWh would be more correct..the rated miles product is also just 2 sig figs, so that is also 3.2kWh.)

So that all seems to align pretty well (1.5% high on the trip meter, though almost certainly due to rounding error on the rmi). For two sig figs it is exactly the same.

This is a super short run, not a good example of how to do this, so the result easily could have been off by 250Wh due to rounding error, but happened not to be. Long drives would be much better so that you could get the results to three significant figures. (You need a rated miles change of greater than 100 miles.)

I noticed that I got out of the car at the grocery at 287 miles, but when I came back to the car it was at 288 miles. So pretty easy to have an error of a mile or two which for short runs would be significant due to this minor jumping of the rmi.

 

[AlanSubie4Life]

A few more datapoints from today's driving:

Ended last night at 280.
Started at 276 (lost 4 miles overnight, hooray, this is normal!).

5 driving segments today:
276rmi -> 274rmi
4.4mi @ 95Wh/mi

0.484kWh vs. 0.418kWh

274rmi -> 263rmi
10.7mi @ 235 Wh/mi

2.66kWh vs. 2.51kWh

263rmi -> 249rmi
9.8mi @ 336Wh/mi

3.39kWh vs. 3.29kWh

248rmi -> 228rmi
19.2mi @ 236Wh/mi

4.84kWh vs. 4.53kWh

228rmi -> 198rmi
19.5mi @ 359 Wh/mi

7.26kWh vs. 7kWh

Total: 18.6kWh (77 rated miles assuming 242Wh/rmi) vs. 17.8kWh (trip meter), error is 4.5%

So, the trip meter did appear to read consistently low, with the 242Wh/rmi assumption (by ~5% in this case).
I wonder whether the error would be similar for a much longer single trip.

Overall stats (since last charge):

Rated miles: 293-198 = 95 rated miles. However, 5 of those miles were used when I was not in the car. So 90 rated miles used while driving (above segments in this and prior post total to 90 rated miles).
Rated miles to kWh -> 90rmi * 242Wh/rmi = 21.8kWh
Since last charge meter: 73.1mi @ 286Wh/mi = 20.9kWh.
(So, 21.8kWh (rmi) vs. 20.9kWh (meter), off by 4%).

Looking at overall efficiency: 95 rmi (23kWh) vs. trip meter of 20.9kWh. (But this isn't a relevant comparison because it includes idle losses.) But those losses will make the rated miles consumption and additional 5% higher than the "since last charge" meter...and those 5 miles lost are very real...for a total of 10% error (23/20.9) vs. that meter.

As previously discussed, possible explanations for the 4-5% error:
1) One explanation is the constant - perhaps the "correct" constant is 230Wh/rmi, that would give perfect alignment, but would imply usable capacity of a new battery is lower than 75kWh (71.3kWh). So I doubt this.
2) Errors due to rounding & state of charge estimation errors. A longer trip would definitely be a better test case. My longer trip that I did gave better alignment than this (it was within a couple %).
3) The trip meter is inaccurate and gives a couple % better efficiency that you're really getting. (It's off by no more than 5%.) (As indicated in the other thread, it could be unmetered energy use due to battery heating, inaccuracies in the monitoring of power consumption, etc.)

Could be a combination of these factors.

Hopefully your test results align pretty well with the above. Your original concern seems valid, though I think the magnitude of the error is probably smaller than you initially thought.