Battery degradation driven by aging, charge cycles and charge current

[Oleksiy]

Here's the paper I'd like to share and discuss: https://www.researchgate.net/publication/286242928_Lifetime_Analyses_of_Lithium-Ion_EV_Batteries. I stumbled upon it while researching the best current to charge my car at to promote the battery's longevity, that's upon purchase of OpenEVSE charger recently. Sorry if it's been covered before.

The results appeared quite surprising and counterintuitive to me, to summarize, they considered three major battery degradation drivers:

(1) calendar aging
(2) charging cycles wear (driving related), and
(3) charge current related aging

For (1) and (2) they used three temperature levels for the analysis (battery temperature, not ambient one) - 40C, 25C and 5C (104F, 77F, 50F).

To analyze (2) they used three cycling patterns all with depth of discharge at 25% - up to 50% (low), to 75% (medium) and up to 100% (high) of SOC. No wonder the lowest degradation level was achieved at the low SOC cycling with the temperature effect quite predictable - the higher the temperature - the faster the degradation. What is surprising, the medium and high SOC cycles show significantly higher degradation at lower battery temperature, and this is especially pronounced for high SOC cycling. So, if we charge to 75% and beyond, degradation of the battery will be faster when its temperature is 10C than in case of high 40C temperature. 25С temperature seems to be the sweet spot, providing for the least degradation at any SOC cycling except for the low SOC where it's on par with 10C temperature level.

Considering this, I think that small batteries (like in my i3 ) stand no chance in cold climates (like in Ukraine ). They are forced to be charged to 100% every day just to deliver the needed range in most cases, and that's exactly the best way to destroy the battery as fast as possible.

On (1), the results were no less unexpected, from my perspective. Overall, EV storage at high temperature is positively correlated with faster degradation at all times, but, judging by the chart they provide, in case of 25C and 10C battery temperature, the battery capacity decrease seems to peak at 60% SOC and then slightly decrease up to and including 100% SOC. Also, it looks like the lower SOC (and voltage) the battery is stored at, the better it is for its long term health. These results contradict the otherwise well established common wisdom on li-ion battery storage. And this is both true for all manufacturers out there except BMW (store at medium level SOC) and BMW (always be charging - store at 100%).

On (3) they say that it's best to avoid lithium metal plating of the anode's surface, because this results in effectively quick death of the battery - accelerated degradation ( (usually occuring at 80% SOH and below). Such plating is driven by high current charging under low battery temperature. They tried charging the cells at 1C, 0.5C and 0.2C rate for a period of 18 months. In case of 1C charging rate the battery dies inevitably, the process is launched immediately. The lower 0.5C charging rate delays the plating development up until about 550-570 cycles, after which we observe accelerated degradation as well, similarly to the 1C test. And the lowest 0.2C rate doesn't seem to trigger any additional degradation aside from normal calendar and charge cycle aging of (1) and (2), at least for the experiment duration.

 

[Anonymous]

Wow, thanks for posting this. Very interesting, and like you say, counter-intuitive, as a lot of the anecdotal real-world evidence points to heat as a major culprit in premature battery degradation. A lot of older Leaf's with poor to non-existent battery management software in AZ are prime examples.
I wonder how much (if any) of their "adaptive charging" to increase battery longevity parameters are used in the BMW battery management software?

 

[Oleksiy]

I think, all producers do implement strategies to limit capacity degradation in the BMS as much as they can. E.g., Leafs significantly decrease DC charging current when the pack is either too hot or too cold. Only, in most cases OEMs seek some compromise between the long term - health of the battery, and short term and more important stuff - user experience, there are lots of trade offs. That's why BMW allowed a scary 2+ DC C-rate for its first i3 (up to 50 kW?), and also stated in the manual that we should charge the car whenever we have an opportunity to, preferably by L2 charger, and store it at 100%. Note that all of these are detrimental for the health of the battery, we should do exactly the opposite to extend its lifetime and number of cycles.

As to the anecdotal evidence, there's obviously a lot more of it in the Leaf community. Leaf owners are no battery degradation deniers, which is the case for many of i3 folks, and they can follow SOH closely both via dashboard bars and using LeafSpy. We don't have this luxury, and even if we check Batt.Kapa.max, it can go up and down randomly thus undermining its validity. Extreme heat is the main culprit in the capacity loss, no doubt about it. But local Ukrainian Leaf owners (and Russian as well, in the far eastern part of their country) report significant capacity degradation after our often frosty winters (AFAIK, Leaf BMS stops updating SOH number under low temperature until it gets warm outside, thus presenting owners with a nasty surprise in spring). I believe, the conclusions of the research report explain this phenomenon as well.

 

[viking79]

Cold is often ignored, but is much worse than high heat for battery degradation, however it is easy to heat a battery. The i3 will heat the battery when necessary and plugged in, so will most EVs. PHEVs can minimize battery use when the battery is too cold. It is a good reason to leave your EV plugged in in winter.

jes.ecsdl.org has some great battery studies as well, looking at regenerative braking impacts, etc.

 

[Oleksiy]

My i3 will heat up the battery only up to 10-11C during the overnight charging and preconditioning. I add 2 more on my way to office where the car is parked for the whole day outside. When I check the battery’s temperature at home right upon arrival, it’s 6-7C. And that’s now, when the ambient temperature is just below freezing. I’ll check if things deteriorate once it gets colder in the winter.

Anyways, based on this paper, the battery seems to operate at a temperature too low to preclude accelerated degradation. No wonder that my i3, that spent its time in Bridgeport, not in Houston, before I brought it to Ukraine, is in mid 17 kWh now, just after 22K miles.

 

[gt1]

Thanks for posting the study. Not an unexpected conclusion. I have a PHEV witha 30 miles battery, which means a full cycle almost every day. I don't expect the battery to last as long as in Tesla or Bolt and I don't think anything can be done about it.

When is the best time to charge in the lower temperatures- right after the arrival, before the battery cools off or before the departure to drive off with a warm battery?

 

[i3Houston]

When is the best time to charge in the lower temperatures- right after the arrival, before the battery cools off or before the departure to drive off with a warm battery?

I believe that is the best time and thats what I do too;

 

[alohart]

In trying to assess the applicability of this research published in 2015 to the battery cells used in the i3, I would expect the conclusions to be valid but the exact details to vary somewhat because battery cell manufacturers have been continually modifying their cells to reduce the rate of capacity degradation.

I consider this research to validate my belief that BMW's recommendations for battery charging ("always be charging") and storage (store at 100% charge level with EVSE connected) increase the rate of capacity degradation but make i3 ownership more convenient and maximize current available range. i3 owners who plan to keep their i3's past the battery pack warranty expiration and who want to minimize capacity degradation should ignore BMW's recommendations and follow the recommendations of this research:

(1) Calendar Aging
"The SoC influences the calendar aging in a nonlinear characteristic: SoC values of 60% and above have caused much higher capacity fades than SoC values below 50%."

I.e., to minimize capacity degradation during storage, store at charge levels <50% and at temperatures near 25º C (77º F).

(2) Aging by Driving Operation
"To optimize the lifetime of a lithium-ion traction battery, it is beneficial to operate the battery at low and medium SoC. A full charging of the battery should only be performed, when the maximum driving range is really required. Charging the battery right before driving can also help to reduce the times of high SoC, in which increased calendar aging leads to an accelerated capacity fade."

The low, medium, and high 25% charge level ranges tested were from 20% to 45%, from 42% to 67%, and from 62% to 87%, respectively. Note for a 60 Ah i3, an indicated 0% charge level is an actual 10%, and a 100% charge level is an actual 95% due to charge level buffers at the low and high charge level ranges that cannot be used. So even the high charge level range test does not start at an indicated 100% charge level, yet the capacity degradation rate when operating in the high charge level range was faster than when operating in the low and medium charge level ranges.

(3) Aging by Charging
"For the cell charged with the maximum rate of 1C, nonlinear aging seems to start already from the beginning, whereas for the minimum rate of 0.2C, the phenomenon is not observed at all within the experiment duration. Reducing the charging rate from 1C to 0.5C, the occurrence of nonlinear aging can be delayed for approximately 500 [effective full cycles]."

A 1C charge rate would charge an i3's battery pack from empty to full in 1 hour which would correspond to DC fast charging. No capacity degradation was observed when charging at 0.2 C which would charge an i3 battery pack from empty to full in 5 hours. AC charging is slightly faster than this charging rate, especially for 60 Ah battery packs. So minimize the frequency of DC fast charging to minimize capacity degradation.

 

[gt1]

Samsung says that 94AH cells should last over 4000 cycles, which corresponds to 500k miles. https://pushevs.com/2018/04/05/samsung-sdi-94-ah-battery-cell-full-specifications/ The 60AH cells shouldn't be that different. I hope it eases the concerns.

 

[Oleksiy]

The low, medium, and high 25% charge level ranges tested were from 20% to 45%, from 42% to 67%, and from 62% to 87%, respectively. Note for a 60 Ah i3, an indicated 0% charge level is an actual 10%, and a 100% charge level is an actual 95% due to charge level buffers at the low and high charge level ranges that cannot be used. So even the high charge level range test does not start at an indicated 100% charge level, yet the capacity degradation rate when operating in the high charge level range was faster than when operating in the low and medium charge level ranges.

Art, how did you see the charge levels were specifically these (62% to 87% and below)? Based on this chart https://www.researchgate.net/profile/Simon_Schuster/publication/286242928/figure/fig2/AS:304417787138050@1449590264515/Cycling-windows-for-the-investigation-of-aging-caused-by-the-driving-operation-8.png, their top cycling window was limited at 4.1V, which corresponds to 100% charge (suggest disregarding the buffers - all producers have these, and to charge to the actual "100%" you need to go up to 4.15V or even 4.2V per cell, not 4.1V, and the BMS wouldn't let you do this anyways).

 

[CanisLupus]

I've read the paper and will continue to use it as a reference . From now till usually about late March , our temperatures are below 10C . The paper suggests that I should use the 20-45 % SOC area unless more range is needed . Considering our buffers that means a 12-42 % reading approximately . Both operation and storage degradations are minimized , I think . Thank you Oleksiy for posting this .

 

[Kristian]

If a lower state of charge is recommended unless extra range required, how can you warm the battery without increasing the SOC?
Is driving with a lower SOC and a cold battery better than a higher SOC with a warm battery?

 

[TheMK]

If a lower state of charge is recommended unless extra range required, how can you warm the battery without increasing the SOC?
Is driving with a lower SOC and a cold battery better than a higher SOC with a warm battery?

I live on top of a hill (some people in other parts of the country would call it a mountain), and so I can easily pick up ~5% SOC in the first 5-10 minutes of my drive. For this reason charging to 100% backfires, because the car will not regen, and then you have to use the friction brakes. On my ICE car I had to replace the brakes far more often then most people do.

What I do is I set up the time of day rate meter to charge from 12:00am until whatever time I set it to. Using the reduced 240V charging setting, it adds about 20% SOC per hour of charging, and I try to target 80% SOC. For example, if you ended the day at 40% SOC, charge from 12am-2am to get to 80%. Then, I press the "climatize now" button on the BMW Connected app on my phone. While this will restart the charging cycle, it also heats the cabin (and I would presume the batteries as well). I usually press that button about 10-20 minutes before I want to drive, and so then when I get out there, the car usually has about 82-83% on it. I unplug, and then drive away. Doing this has allowed me to get the full regen rate while driving down into town, and will preserve my brakes from needing frequent replacement.

 

[jadnashuanh]

FWIW, setting a departure time's primary function is to warm the batteries. IT triggers a 1Kw electric heater to start to warm the batteries. On that menu, you have a chance to warm the cabin as well. During this process, the car could be pulling over 20A at peak. Preconditioning the cabin does not directly heat the batteries, so to get best range on the i3, it's good to do that.. Not wanting to try to overcharge the batteries is also one reason why you can't start the REx with a higher SOC than 75% - there must be some place for that energy to go. Assuming the guy who wrote the i3 E-book knows what he's talking about, charging the car once from 0-100% puts the same wear on it at doing it 10x from 90-100%. After over 4-years, mine seems to have retained within 95-97% the original battery capacity. When I get home, I plug it in and leave it, often setting a departure time in the winter, but not when it's warmer where I'll wait to precondition the cabin. You can still set a departure time, but unless you set it to preconditioning the cabin, it won't perform any battery warming. Warming the battery allows full use, otherwise, it could be restricted until they warm up some through use. Warming the batteries also helps when charging, as cold batteries won't take as fast of a charge.

 

[Oleksiy]

If a lower state of charge is recommended unless extra range required, how can you warm the battery without increasing the SOC?

What I did before I chose to put the car in a garage for the winter, I set the off-peak time setting and the departure and preconditioning time for the period needed for the battery to get to about 90% from the start of the charging (note that in case the BMS sees that the off-peak time slot is not sufficient to get to 100% SOC, it will start the session immediately regardless of the start time settings, so you may want to plug it in once your off-peak rate begins, or alternatively decrease the amperage accordingly). In the morning the car was at about 80-85% with the battery temperature of 10-11C (50-52F). In some cases when I needed full battery capacity and/or let the car balance the cells overnight, the preconditioning outcome pack's temperature was the same, albeit at 100% SOC.

Is driving with a lower SOC and a cold battery better than a higher SOC with a warm battery?

Warm battery (especially out of a warm garage in the morning) dramatically increases expected mileage in my case under subfreezing ambient temperature - the car is quite close to its mild weather efficiency. So, in winter I prefer to have the battery warmed up or preconditioned at any time. Also, (relatively) warm battery is healthier for the recuperation charging - Li-ion degrades very fast if you charge it at high C-rates in cold state. The BMS should limit the regenerative braking intensity accordingly (and it feels like it's not as efficient as in summer indeed), but I'm not sure whether it's properly calibrated to extend the life of the pack. The car still pushes energy back into the battery fast - I observed the efficiency of up to 2.5 km/kWh yesterday (the pack's temperature on my way home in the evening was at 0C or 32F) vs. the summer peaks of about 1.5 km/Kwh.

 

[Kristian]

I've tried to only partially charge the battery before (I use a timer on the socket to control when it starts) with my trickle charger set to 2.4kw and after 3 hours the battery will be nearly fully charged 80% to 100% and the battery will only have increased in temperature by a couple of degrees - I'll have to test it again.

 

[flyboy320]

How are you measuring the battery temperature?

 

[Kristian]

How are you measuring the battery temperature?

Using the hidden menu (press and hold the mileage trip reset for 10 seconds, scroll down to unlock, type in sum of 1st 6 or 8 digit of vin, scroll to temperature menu)
There are better instructions possibly on this forum for unlocking the menus

 

[flyboy320]

Using the hidden menu (press and hold the mileage trip reset for 10 seconds, scroll down to unlock, type in sum of 1st 6 or 8 digit of vin, scroll to temperature menu)
There are better instructions possibly on this forum for unlocking the menus

Ah, it's in that menu, I've just used that for Kapa.Max, didn't know it also had the battery temp in there as well, thanks.

 

[Oleksiy]

I've tried to only partially charge the battery before (I use a timer on the socket to control when it starts) with my trickle charger set to 2.4kw and after 3 hours the battery will be nearly fully charged 80% to 100% and the battery will only have increased in temperature by a couple of degrees - I'll have to test it again.

If you did tick the preconditioning box as well, the battery should have been warmed up to 10-11C. The system just won't warm it any further (at least according to my observations), so the temperature gain may depend on your starting temperature as well.