Capacity fade in EVs

[GetOffYourGas]

I also monitor the energy used to recharge the car. Charging station logs it. So far, at least, that information is consistent with the LeafSpy estimates. Charged a few days ago from 35% SOC to full using 14.7kWh. 240V 30A L2 station. I compute about 92% of capacity.

Note that with LeafSpy full is actually 95% SOC, and stop is at about 2% SOC. The dash meter for state of charge reads 100% when LeafSpy reads 95%.

Now, if I was bored, I'd drive the car until turtle, park by a charging station and run the heat until shutdown. Then record the energy to recharge the car, and compare with the 25.4kWh needed to recharge the battery from empty to full when new. I'm not bored, yet.

Honest question - do we know that charging efficiency is more or less constant as a battery degrades? Or does a degraded battery have more internal resistance, and therefore less efficient charge cycling? If the efficiency goes down, then the battery would look less degraded than it is, since it would take more energy to charge.

Again, not a critique, but an honest question.

I have done nearly zero testing of my battery. My dash reports 11 or 12 capacity bars (roughly 79-85% or original capacity) after 5 years / 41k miles. I have tried to take care of the battery. I charge to 80% most of the time. If the battery reads 6+ temperature bars, I don't charge the car unless absolutely necessary. Fortunately where I am, even in the summer, the battery cools off to 5 bars most nights.

Given what I knew about batteries in 2012, I expected Nissan's estimates of 80% at 5 years / 70% at 8 years to be far too optimistic. In hot parts of the country I was right. But here, I seem to be tracking that.

 

[WetEV]

Honest question - do we know that charging efficiency is more or less constant as a battery degrades? Or does a degraded battery have more internal resistance, and therefore less efficient charge cycling? If the efficiency goes down, then the battery would look less degraded than it is, since it would take more energy to charge.

Yes, the charging efficiency would degrade due to rising internal resistance, but this should be a much smaller effect than capacity loss as long as the charge rate is well less than C. Leaf with 6kW charger/24kW is C/4. Fast charge, C>1, would be more problematic. Loss in internal resistance is I^2*R, so halving the current would quarter the loss.

Given what I knew about batteries in 2012, I expected Nissan's estimates of 80% at 5 years / 70% at 8 years to be far too optimistic. In hot parts of the country I was right. But here, I seem to be tracking that.

Glad to hear this.

In cool coastal Pacific Northwest, Leafs I know about seem to generally be doing slightly better than Nissan's estimates, with one exception. While I can see why a TMS is a gain for battery life in Arizona, I doubt if it is in the north in general, and judging by a few examples with TMS, might even be a drawback in the PNW. A TMS keeps the battery warmer in cool conditions, which for all things being equal, should reduce battery life relative to a passive cooled BEV.

 

[oilerlord]

We bought because we keep our cars until they die of old age. So I expect our Bolt to be with us for very many years. I'm not worried about the batteries too much after reading this
http://www.teslacentral.com/worried-about-tesla-battery-degradation-its-23-miles-every-100000-driven

A lot of people buy cars and drive them until the die, but you and I are in a very small minority attempting this with an EV. In terms of helping the environment, not only are you not burning gasoline & changing oil every 5000 miles, you aren't wasting the earth's energy & resources replacing a perfectly good car every three years. On behalf of Mother Nature...Thank you.

The truth about battery degradation is probably somewhere in the middle of this conversation. Batteries degrade, but the rate they degrade is also mitigated by how the owner maintains it. Just like hypermilers that go on forums boasting that they beat the EPA mileage number by 40%, I'm sure there are EV owners with OCD that make it their life's mission to care for their car's battery. It's those owners that go on forums that then make claims that their battery didn't degrade over 100,000 miles. I'm sure if I kept my EV parked at our house in Arizona in the heat of a couple of 100 degree summers at 100% SOC - my battery wouldn't be so lucky.

 

[Lumpy12]

One other variable will be the price of replacement batteries when those in our Bolt do reach "end of life." I expect there will be used replacements available, and also new from GM.

With any car, you take a gamble on buying it. Is it a lemon. Should I buy or lease. What's it's resale value going to be. We made our best guess. The environmental factor was a big one for me as I'm a true tree hugger :mrgreen:

 

[oilerlord]

If you're going to keep the car until it dies of old age, don't worry about resale value because it won't have much. I've mused about the possibility of replacing my car's battery after 100,000 miles but it probably won't come cheap. I might consider a DIY refurbishment using a pack from a wrecked Model S, but its more likely that I'll just sell it as a 60 mile car for around $3,000. Whatever money I get for it will be gravy given cheap kWh's and not having to pay for gasoline and ICE-related maintenance.

 

[Anonymous]

Hi,

This is a graph of the battery fade of my 2014 Spark EV. It went from about 19.5kWh to 16.5kWh over a period of 3 years and 22,000 miles.

This was the A123 battery that was in the 2014, Chevy changed to LG as a supplier in 2015 - hopefully they are better both in the Spark and the Bolt.

I estimated the capacity from the readings on the Energy Screen following discharges of more than 50%.

kevin

 

[michael]

Your plot is well done and very believable. My Focus with LG batteries experienced a comparable amount of fade over three years, although in my case over 50,000+ miles

People like to believe battery fade isn't an important issue, but I am convinced it is.

Where were you located? What attempts to mitigate fade did you take, if any?

 

[GetOffYourGas]

Great plot, Kevin93. Thanks for sharing.

I see you plotted the independent axis against miles. This assumes that cycling loss is the driving factor (it may be, I don't know). Could you share another plot against time? That would show the curve if calendar loss is the driving factor. Of course, it will probably be similar assuming your driving patterns are consistent.

 

[boltage]

Great plot, Kevin93. Thanks for sharing.

I see you plotted the independent axis against miles. This assumes that cycling loss is the driving factor (it may be, I don't know). Could you share another plot against time? That would show the curve if calendar loss is the driving factor. Of course, it will probably be similar assuming your driving patterns are consistent.

Would the number of charge cycles, and how "full" one charges the battery each time, be one of the larger factors in battery fade?

I.e. someone who drives to 70% and then uses a DC quick charger to top it off to 100% on a regular basis may have different battery fade than someone who drives to 20% and then uses L2 to fill it back to 80% or 90%.

 

[GetOffYourGas]

Great plot, Kevin93. Thanks for sharing.

I see you plotted the independent axis against miles. This assumes that cycling loss is the driving factor (it may be, I don't know). Could you share another plot against time? That would show the curve if calendar loss is the driving factor. Of course, it will probably be similar assuming your driving patterns are consistent.

Would the number of charge cycles, and how "full" one charges the battery each time, be one of the larger factors in battery fade?

I.e. someone who drives to 70% and then uses a DC quick charger to top it off to 100% on a regular basis may have different battery fade than someone who drives to 20% and then uses L2 to fill it back to 80% or 90%.

Sure, there are a multitude of variables that affect capacity loss. I was simply requesting a slightly different graph based on the data he presumably has. I assume that he knows the dates of each reading as well as the odometer.

I wish I had a record of my Leaf's battery. I intend to keep driving it until the *right* replacement BEV comes along, at the right price (the Bolt is close, but still a little steep price-wise). I may keep the Leaf anyway as a winter beater.

 

[Anonymous]

Great plot, Kevin93. Thanks for sharing.

I see you plotted the independent axis against miles. This assumes that cycling loss is the driving factor (it may be, I don't know). Could you share another plot against time? That would show the curve if calendar loss is the driving factor. Of course, it will probably be similar assuming your driving patterns are consistent.

Here is the same data plotted vs Time. Not much different although seems a bit straighter, i.e. a more linear relationship.

I'm in the bay area so the car was not subjected to large temperature ranges. I only discharged it very low two times, mostly it was recharged when it was ~60% SOC. It would sit a few hours at full SOC in the daytime. I used Level 2 charging mainly (3.3kW) - I didn't have the FC option.

I averaged about 5.7mi/kWh (as shown on the energy screen - 4.7mi/kWh for AC input).

kevin

 

[oilerlord]

Kevin,

It sounds like you were kind to your battery:

- You stored / drove the car in a cool environment
- You (more or less) charged the pack in the middle of it's range
- You didn't store the car at full SOC for more than a few hours at a time
- You avoided the use of DCFC (since your car wasn't equipped with that option)

And yet, your range still degraded by ~10% after only 22,000 miles.

I've read that the total capacity for the 2014 model is approximately 21.4 kWh. You started with 19.5 kWh - which I assume is "usable" capacity. That indicates a buffer of only about 10%, so at 100% SOC - you're fully charging 90% of the battery if you choose to or not. Perhaps that's the biggest reason for the somewhat moderate degradation you've indicated.

My car isn't equipped with the "range extender" option that essentially cycles deeper (about 33 kWh) into it's 28 kWh usable / 36 kWh total capacity. My owner's manual discourages the use of that option as it degrades the life of the battery. It looks like the 2014 Spark EV is, comparatively, in "range extender" mode by design.

What do you think?

 

[boltage]

- You (more or less) charged the pack in the middle of it's range

It looks like he charged from 60-100% of nominal capacity (55-91% of actual capacity, if actual capacity is 21.4kWh with nominal capacity of 19.5kWh), so the battery pack was mostly in the high end of the range.

mostly it was recharged when it was ~60% SOC.

 

[oilerlord]

So, given these results (from the OP's link):

I don't believe that Tesla has made a breakthrough and/or is more advanced than other EV's in terms of reduced capacity fade - rather, it appears that the combination of fewer charge cycles along with a larger battery buffer are key to the long term health of the battery. Degradation still happens, but a larger buffer masks it's effect.

Has anyone found out what the total/usable battery capacity of the Bolt EV is, or are we still guessing?

 

[DucRider]

I don't believe that Tesla has made a breakthrough and/or is more advanced than other EV's in terms of reduced capacity fade - rather, it appears that the combination of fewer charge cycles along with a larger battery buffer are key to the long term health of the battery. Degradation still happens, but a larger buffer masks it's effect.

Has anyone found out what the total/usable battery capacity of the Bolt EV is, or are we still guessing?

Tesla Battery buffers are not larger (in either kWh or % of total capacity).
Excluding the models with software limited capacity, Tesla runs between 2.5 an 4 kWh of buffer, or ~3% to 5%.

There are so many variables and unknowns with the Bolt EV battery that capacity fade is impossible to predict. It's going to take a couple of years of real world usage to get a feel for how they are lasting.

 

[oilerlord]

Tesla Battery buffers are not larger (in either kWh or % of total capacity).
Excluding the models with software limited capacity, Tesla runs between 2.5 an 4 kWh of buffer, or ~3% to 5%.

I did not know that.

It does appear that (generally speaking) the software limited models fare better long term than ones that run close to their total capacity:

Either the OP's link is "happy talk" as Michael suggested, or the data provided is skewed coming from Tesla owners that are very careful keeping their SOC in the optimum range. I'd guess that the majority of Tesla owners would just keep their cars plugged in all the time, so now I'm a little puzzled as I had the understanding that a larger buffer was key to masking degradation. Case in point: Volt V1.0 with it's approximate 10.6 kWh usable / 15 kWh total battery capacity. Many of those owners report zero (noticeable) degradation after 100K miles.

How would you explain this, Gary...fewer charging cycles (i.e. having a larger battery) is more important than buffer size? All things considered, I believe that the Spark EV has excellent BMS/TMS. Given what we know about Kevin's 2014 Spark EV, why would it lose 10% of it's capacity after only 22K miles?

Edit: Found this humorous...Tesla charging nearly $10K for a "range upgrade" that is nothing more than a software update.

https://electrek.co/2016/12/14/tesla-battery-capacity/

 

[michael]

Remember also that Tesla, unlike most EV makers, provides a "slider" to limit the maximum charge level. If it's set at a low level, leaving the car plugged in has less harmful effect that in a Leaf, Focus, Spark, etc. that always charge up to the maximum allowed level.

And my reading seems to support my understanding that these Tesla degradation estimates are based on the "rated range" dash display.

For example, one writer says:

The way to measure this is to do a full charge (100%) and then check the EPA rated range (in North America) or Typical range (in Europe and Asia/Pacific). In the plot, these numbers are then compared to the range numbers the car displayed when it was new. For example, for the 85 kWh Model S85 variant, this is about 400 km typical range or 265 mi EPA rated range. Even though this is mostly a reliable method, sometimes the computer in the car can’t accurately estimate how much energy the battery holds and might display an inaccurate range number. To improve accuracy, it is a good idea to run down the battery to almost empty and then charge to 100%, once a month. This is known as rebalancing the battery. However, the battery shouldn’t be left at 0% or 100% for more than 2 hours.

As he acknowledges, "sometimes the computer can't accurately..."

Until someone can demonstrate that the Tesla dash displays of battery capacity are accurate (I know the Ford ones are NOT) then I need to regard them as unproven.

If someone is serious about predicting Tesla battery fade, I think the way to do this is to buy some Tesla cells (they are readily available on Ebay) and measure the degradation of a sample of individual cells, at various temperatures, through different charging profiles (full, partial, etc)


I know there are numerous studies out there that try to address this, but I haven't yet found one that uses Panasonic NCA cells (similar or possibly identical to those used by Tesla) and gives a clear answer. I'm still looking, maybe someone else can find a study that ages these cells over, for example 1000 cycles, 50% DOD, one cycle per day, at some defined temperature (maybe 40 C?) I think this would reasonably simulate 100 miles a day in a Tesla for three years, 100,000 miles. That's something I would like to see and compare with the Tesla Fan-boy analysis.

 

[oilerlord]

Right, I forgot about the charge level slider in the Tesla. From what I'm reading, the default charge level is set to 90% (not sure if that's based on total or usable capacity). Hard to say if the list of results from the OP's link were charge levels set to default or some other user-selected SOC limit.

I'll continue to limit the charge maximum and endeavor to not keep my battery at 100% (indicated) SOC for more than an hour or two. The latter is pretty easy for me, but when I'm coping with 40-50 miles of range when it's -15F - the former isn't so easy.

 

[michael]

I found an interesting study

http://www.myrav4ev.com/forum/viewtopic.php?f=8&t=561&p=13780&hilit=battery+fade#p13780

published on the Rav 4 EV forum.

The author purchased Panasonic cells similar or possibly identical to those used by Tesla, which are presumably also the same or similar to those used in your Mercedes.

He tested the cells in isolation, at room temperature, and subjected them to cycling equivalent to approximately 90K miles. He said this took "many months"

In his tests, fade was about 10% at 90K miles. However, these were at moderate temperatures, and they involved primarily cycling fade, not calendar fade. Typically, a 100K mile car would need 6 to 10 years, depending on usage.

If his tests are indicative, it tends to dispute the often quoted Tesla owner report that suggests about 10% per 100K miles, even with temperature and calendar effects.

Another interesting study

http://jes.ecsdl.org/content/164/1/A6066.full#F6

From the Journal of the Electrochemical Society looks at calendar fade in NCA batteries, similar to those used by Tesla. they basically charged up cells to various states of charge and stored them, at various temperatures, for 9 months. They then measured the battery capacity when fully charged.

Look in particular at Figure 8. It shows, as expected, greater fade at higher storage temperatures and states of charge. But below about 55% SOC, fade is very low and almost independent of SOC. Above 55 %, there is a step increase in fade rate, and the fade increases with higher and higher SOC.

This study gives results which are consistent with an earlier Army study which showed that the cells lasted better when cycled between 0% and 50% compared to cells cycled between 25% and 75% or between 50% and 100%. Many people believe that "mid level" is best for the batteries, but these two studies refute that belief. Lower is better than midway.

 

[oilerlord]

Hey, thanks Michael. Great info!

I suppose that Mercedes asked Tesla to reduce the usable capacity of my 36 kWh battery down to 28 kWh for the purpose of offering a "range extender" option they could charge extra for - much like Tesla offers their customers a "range upgrade".

For what it's worth, I'm happy that (by design) I'm only able to charge up to 80% of my total capacity so that should help minimize battery fade somewhat. Most of the advice I get from other B250e owners is to simply not worry about it - mostly because of that 20% buffer. I may explore / research charging to 75% (using 21 kWh out of the total 36 kWh) and running it closer to zero as that's more in line with the Army study you're referring to. I go back to the example of Volt V1.0 and it's large buffer. I tend to believe there is a correlation between cycling 0-50% total capacity and original Volt owner's reports of little battery degradation after 100K miles.

Thanks again.