EV Li-ion battery lost 10% while parked for 4 hours

[keepgoing]

2014 BEV Giga model. This morning I parked my car after dropping off kids to school (2 mile round trip) with about 80% battery level left. When I got into the car 4 hours later, the battery level is at 68%. The weather is around 40 degree F and I drove with Eco Pro + unless weather dropped to freezing. No pre-heat or "Climatize Now" operation either, as far as I know. I did not have any headlights on either, just the DRL while car is being driven. Anyone notice the battery level self adjust this way and this much?

I also noticed a while back sitting at traffic light that the battery level would drop in 1% increment for a total of about 3% just sitting at the light (<1min). Again, Eco Pro + and no climate control was being used. I was not using the brake to keep the car still at that moment since it is stopped by itself at a relatively level road, may be a little incline up. I was thinking may be the motor was actually using electricity to keep the car from sliding back.

It is kind of big deal in planning trip since the GOM now only estimate 50 miles of range for comfort mode and 60 for ecopro+. Every mile counts, compare to 4 years ago when the GOM estimate of 80 miles for comfort and 90 for ecopro+.

 

[gt1]

You can try to setup ISTA diagnostic software and monitor the cell voltages with it. This way you'll know if your battery voltage really drops or the % readout is having issues.

 

[eNate]

I will only add the comment that, yes, electricity is used to hold the vehicle from rolling backward (though 3% seems high), and that determining state of charge -- especially in the middle range -- is complex to calculate and not precise.

I think you have to look at your range as a whole and see what your battery does over the course of a long drive. It seems that standing still your system is detecting some "electrical leakage" and it's making adjustments to your SOC accordingly.

 

[symonray]

On occasion, I have seen the SOC on my 2016 i3 Rex drop anywhere from 2% to 5% after sitting in a cold parking garage for 8 to 12 hours during the winter months. I attributed the SOC loss to the loss of capacity caused by the cold temperatures while the vehicle is sitting idle.

 

[eNate]

... I attributed the SOC loss to the loss of capacity caused by the cold temperatures while the vehicle is sitting idle.

That's another good example of why SOC estimates are tricky.

When a battery is cooled, it becomes more stable, i.e lower self-discharge. So parked at a low temperature shouldn't have much impact to actual state of charge.

But then when that same battery is put into use, because it's cold, the reaction taking place inside is less efficient, so we get fewer miles out of it.

I have no clue what algorithms BMW uses to estimate SOC, but voltage, temperature, and tabulation are all almost certainly in use. So for a parked i3, as cell temperature drops, the software is probably also measuring voltage, and via a lookup table, estimating that at a specific temperature, such-and-such a voltage indicates a pre-determined SOC.

But the system **should** (assumption on my part here, based on the state of current battery technology) also be constantly modifying the tables as the battery ages, knowing that if the car is parked at 50% in the afternoon, and the temperature dips 20° F , that the "cold voltage" it reads in the morning is the new 50%.

Just as important is the "tabulation" function (Coulomb counting) where the current in and out is being very carefully tracked, and the system estimates SOC based on [Starting SOC] - [mA out] + [mA in]. This would most likely be what is primarily responsible for our SOC estimates as we drive our i3's, because it's impossible to accurately measure voltage with the degree of precision necessary while the battery is in use. The problem with this method is exactly what you'd expect -- that it's subject to error, and can really only "reset" when the battery is full (ours are never truly at 100%) , and, as before, needs to learn to adapt to an aging battery.

At the literal "end of the day," when the car is parked, maybe charged, and the battery temp has returned to ambient, the system must rectify its ledger with the voltage measurement, update tables as necessary, and "decide" what the SOC is.

There's a lot of good material out there discussing the various techniques, some EV-specific, and some get very heady, buried in complex equations I can't follow. If there's one thing I solidly learned from it: estimating battery SOC is a dark art and what your dashboard displays with certainty could easily be off by a few percentage points.

 

[panamamike]

Given all of this I wonder how they estimate the eMPG. If we end up losing energy just to maintain the battery at a correct state, that should be figured into the MPG.

 

[eNate]

The MPG test is standardized over about 20 miles of simulated city and highway driving, but not prolonged idling.

MPGe is based on the same standard "course" (it's actually performed on a dyno), and convert a gallon of gasoline into its electrical equivalent -- about 33 kWh -- to calculate a MPG equivalent.

A dyno can simulate a hill while in motion, but not while stopped (i.e. anti-rollback motor brake electrical usage can't be accounted for), and these tests are normally run with all HVAC / lights / accessories off in all cars.