Battery Degradation DataBase

My Nissan Leaf Forum

Help Support My Nissan Leaf Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

joeriv

Well-known member
Joined
Mar 24, 2013
Messages
287
Location
Fairfield County CT
There is finally a definitive study on battery degradation by Geotab:

“A persistent concern among some EV drivers is the long-term health of the battery. All batteries lose some storage capacity over time. But how might that degradation affect your driving range a few years down the line? To help answer that question, we can now look to Geotab, a leading telematics-fleet-management company with access to a lot of EVs. Lo and behold, the losses are minor.

Geotab created its so-called EV Battery Degradation Tool by pulling data directly from the 6,300 EVs from its fleets. What’s super-cool is the interactive nature of the tool, allowing users to drill down to 21 specific electric models.

You can use the tool to slice and dice the data for yourself.

In June 2018, Geotab acquired FleetCarma, a forerunner in providing technology support the use of EVs in fleets.

Here’s a quick rundown of what the data revealed:

If current degradation rates are maintained, the vast majority of batteries will outlast the usable life of the vehicle.

The average decline in energy storage is 2.3% per year. For a 150-mile EV, you’re likely to lose 17 miles of accessible range after five years.

EV batteries decline in a non-linear fashion. There’s an early drop, but the rate of decline slows down in subsequent years.

Liquid-cooled batteries decline slower than air-cooled packs. Geotab saw that a 2015 Tesla Model S with liquid cooling had an average annual degradation rate of 2.3%, compared to an air-cooled 2015 Nissan Leaf’s rate of 4.2%.

Battery-powered vehicles that have bigger state-of-charge buffers fare better. In other words, some carmakers use a smaller percentage of the battery’s capacity, which reduces usable range. But the conservative approach slows down the degradation rate, most notably in early versions of the Chevrolet Volt plug-in hybrid.

Higher vehicle use does not necessarily equal higher battery degradation.

Vehicles driven in hot temperatures show a faster decline in battery health.

The use of DC fast-chargers speeds up the process of degradation, but there’s not much difference in battery health based on frequent use of Level 1 versus Leve 2 charging. Losses that happen with frequent DC charging are made worse in hot climates.”

https://electrek.co/2019/12/14/8-lessons-about-ev-battery-health-from-6300-electric-cars/

The database tool is here:

https://storage.googleapis.com/geotab-sandbox/ev-battery-degradation/index.html
 
^^^ Inevitably, the results will be skewed by the geographical distribution of the cars at the time. Ideally we'd want to be able to select for that category as well, but this seems to be the best, most objective data we've got at the moment.
 
GRA said:
^^^ Inevitably, the results will be skewed by the geographical distribution of the cars at the time. Ideally we'd want to be able to select for that category as well, but this seems to be the best, most objective data we've got at the moment.

Fleet Karma is a Canadian company. Is there a way to know how much of the data is from that up north colder climate, Eh?
 
If the data is based on how much wall power went into the cars, this is a great study as it would be more objective than GOM figures or driving variabilities.

One concern I have had with most Tesla degradation claims, or lack there of, has been that it’s based on the GOM range on the car, which underlying calc is not a constant. This goes in both directions and it even more opaque as the software has been updated.
 
I went to the FleetCarma website to find out more about this study and found the following:

1. The study consists of cars in Arizona, Rhode Island, New York, Nashville TN, Florida and Lincoln, Nebraska. Other areas may open up later.

https://www.fleetcarma.com/smartcharge/programs/

2. Data is collected from EVs using the FleetCarma C2 OBD II device and downloaded to them via a cellular.

https://www.fleetcarma.com/smartcharge/profile/

3. Apparently they pay users who participate in these studies and users are able to get detailed data from the OBD II device as well a FleetCarma (similar to LeafSpy?). It seems that participation is limited - Arizona for example is limited to 500 cars.

I would love to get the raw Leaf data but I doubt that’s possible without paying for it.
 
This is at least hopeful that the spread is better than originally thought geographically.

If it’s Obdii data, wonder how they are getting the Tesla data as I didn’t think they had an SOH value available in their codes.
 
For the LEAF, this data may be somewhat skewed depending on how many 2018+ cars there are in the sample. If the majority of cars are 2011-2017 that have the older battery design, that might give the impression that the reliability of the LEAFs battery is less than other current makes.
 
OrientExpress said:
For the LEAF, this data may be somewhat skewed depending on how many 2018+ cars there are in the sample. If the majority of cars are 2011-2017 that have the older battery design, that might give the impression that the reliability of the LEAFs battery is less than other current makes.


The is no one "older battery design." It's more like three designs. You've also been telling us that the 30kwh pack is great, but now it's not reliable...?
 
For its generation it has been a good incremental improvement, but has since been superseded by the current generation of battery.

At some point the current generation will have something better replace it as well.
 
It’s got about 22K on it, I keep it between 30 and 80%, and the GOM is pretty accurate if I keep my speed below 70, 145-175 is the usual forecast,. Freeway, ePedal off, ProPilot engaged, typical heavy traffic, gives me the best results.

It’s good enough that I usually don’t even look at it unless we are going up to Napa, Monterey or to the Central Valley.
 
Their 2012 entry looks like maybe one vehicle... and zero degradation after 3 years, so suspicious data also.
2012 vs 2013 (I believe these batteries were substantially similar):
0yUx4RZ.png


2014 & 2015 do look similar:
LAH5pTo.png


2016 & 2017 do look similar:
pqZqkCx.png


...so if you throw-out the bogus 2012 value, then years 2013-2017 show reasonable consistency between "sibling" years. But you get a surprise if you compare a pre-Lizard, Lizard and a gen1 30kw:
TwIzDBF.png


The data makes it look like the 24kwhr Lizard had the worst Leaf degradation, and the 16/17 30kwhr Leaf had the least. This does not match the conventional wisdom.

I wonder if they have controlled for location... for example, if all 2017 Leafs in Phoenix look X% worse than the median 2017 Leafs, then when comparing 2016 vs 2017, adjust any Phoenix samples to compensate. Otherwise, you could have a situation where (say) most of the Leafs in their system from 2014 were in Phoenix, and most of the 2017s were in Toronto, making a 2014 vs 2017 comparison misleading.
 
There are no Leafs in Phoenix or Toronto per my earlier post on this:

“1. The study consists of cars in Arizona, Rhode Island, New York, Nashville TN, Florida and Lincoln, Nebraska. Other areas may open up later.”

The “conventional wisdom” can be misleading at best - see this study backed by data on the Leaf 30 battery:

https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/

I wish we could get the raw data on Leafs in this study but it’s not going to happen.
 
joeriv said:
There are no Leafs in Phoenix or Toronto per my earlier post on this:

“1. The study consists of cars in Arizona, Rhode Island, New York, Nashville TN, Florida and Lincoln, Nebraska. Other areas may open up later.”

The “conventional wisdom” can be misleading at best - see this study backed by data on the Leaf 30 battery:

https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/

I wish we could get the raw data on Leafs in this study but it’s not going to happen.


IIRC that study - and its extrapolations - are based on a grand total of nine cars, and most (all?) of them were not built in North America. Grain of salt suggested.
 
I'm part of Smart Charge here in NY. While there are many awesome things about this program - only one of them being that they pay me 10 cents per kWh to charge between midnight and 8 AM, when the time of use rate is 5 cents per kWh, which means charging my Leaf costs negative 5 cents per kWh - I have very little confidence in the accuracy of their SOH data. For one, I got my Smart Charge C2 device only a week after purchasing my Leaf, and this C2 device said my SOH was 95% (Leaf Spy said around 98%). Since then, my Smart Charge SOH has bounced back and forth between 95% and 90%, but it's only ever these two numbers, no tenths and nothing in between, and it's more often 95% when I'm at a high SOC. The point being that whatever they're using to track the SOH only seems capable of presenting readings in 5% increments. So if that's the data they used for this study, I wouldn't really trust it. I honestly just ignore their SOH readings in my Smart Charge dashboard. Fantastic program otherwise, though. I traded in a 2016 Honda Civic for my 2019 Leaf SV, and thanks to Smart Charge, I'm on track to save around $3,000 a year on fuel and maintenance. By the time my Leaf's battery warranty is up, the car will have literally paid for itself in these savings....
 
joeriv said:
There are no Leafs in Phoenix or Toronto per my earlier post on this:

“1. The study consists of cars in Arizona, Rhode Island, New York, Nashville TN, Florida and Lincoln, Nebraska. Other areas may open up later.”

The “conventional wisdom” can be misleading at best - see this study backed by data on the Leaf 30 battery:

https://flipthefleet.org/2018/30-kwh-nissan-leaf-firmware-update-to-correct-capacity-reporting/

I wish we could get the raw data on Leafs in this study but it’s not going to happen.

That study should be deleted and scrubbed from the internet. It is COMPLETELY WRONG.
 
Back
Top