INL L2 vs DC initial capacity test results after 50k mi+

[edatoakrun]

Unexplained is why the DC charged LEAFs are only losing capacity slightly faster than the L2 Charged LEAFs, but have much larger losses in the 45 mile constant-speed range test.

It's clear to me that the DC packs have increased internal impedance compared to the L2 packs...

...in the I have 28k miles on the car, 53.2 Ah, 64.3 Hx and 81% SOH. 100% Gid count is around 210-215. I've noted my Hx number is is quite a bit lower than others with the same capacity loss...

Which is why I don't think It is very useful to view a single value, a percentage of available capacity (what we are allowed to access) or a percentage of total capacity (what the INL is measuring) as a measurement of "degradation".

All Nissan has said publically is that it warrants a

...LEAF with loss below nine (9) bars (or approximately below 70 percent) as shown on the vehicle’s battery capacity level gauge... it will provide the vehicle with a capacity level of nine bars or more on the battery capacity level gauge...

http://www.mynissanleaf.com/viewtopic.php?f=4&t=13192" onclick="window.open(this.href);return false;

I suspect Nissan was careful enough with it's terms as to avoid more lawsuits, and that ~33.75% gid loss (when the ninth bar is lost) does in fact correspond to a loss of ~30% or less of both total capacity and range, and maybe much less, as was the case with the only 8 capacity bar Leaf range tested in Phoenix in 2012, which was reported to have ~38% gid loss but only ~22% range loss (~59.3 miles at ~62 mph) from the INL test average ranges of "new" LEAFs.

http://electricvehiclewiki.com/Battery_Capacity_Loss" onclick="window.open(this.href);return false;

Unfortunately, I can't recall anyone has posted the results on MNL of either 45, 60 or 70 mph constant-speed range tests, or recharge capacity test results, for any 8 bar LEAF in the nearly two years, since.

Nissan would probably be pleased by that (assuming someone there noticed) as it seems to want the facts about capacity/range loss to remain as vague as the capacity warrantee.

I have considered that, if I lost my ninth bar near 5 years, would want to exercise the warranty, and have Nissan to replace it with a nine or even ten bar used pack, with an unknown use history?

Exactly what assurance do you have that the replacement pack will actually provide more range, further into the future, than the one you traded in?

 

[JeremyW]

Ok, so you're going to ignore all the stuff I said about independent impedance measurements I posted? Only dwell on the numbers I post for reference? Fine...

Regardless of whatever the car thinks is happening to capacity of the pack, the degradation I am seeing is manifesting itself as an increase in internal impedance.

 

[edatoakrun]

Ok, so you're going to ignore all the stuff I said about independent impedance measurements I posted? ...

Assuming you're speaking to me, no I did not ignore any part of your comment, nor did I disagree with you.


I expect the future INL report will have more details on the subject.

Seen the volt battery degradation report?

2011 Chevrolet Volt VIN 0815 Plug-In Hybrid Electric Vehicle Battery Test Results
http://avt.inel.gov/pdf/EREV/battery2011volt0815.pdf" onclick="window.open(this.href);return false;

As discussed in the pages following:

="edatoakrun"]I don't think the AVT test results for the 2011 and 2013 Volt have been posted yet on MNL.

2011 Chevrolet Volt
Advanced Vehicle Testing – Baseline Testing Results

http://avt.inel.gov/pdf/EREV/fact2011chevroletvolt.pdf" onclick="window.open(this.href);return false;

2013 Chevrolet Volt
Advanced Vehicle Testing – Baseline Testing Results

http://avt.inel.gov/pdf/EREV/fact2013chevroletvolt.pdf" onclick="window.open(this.href);return false;

No great revelations I can see, but there is a lot to consider.

Remember to take the lower ambient temperatures for the 2013 tests into account, lest you conclude it was beat out by the 2011 in the 45, 60, and 70 mph constant speed tests.

The thing I'm trying to figure out from those results (if I read them correctly) is why the kWh available and battery roundtrip/overall trip efficiency from both Volt packs decline as the test speeds increase...


If you want to compare to the 2011 LEAF tests, (at a few degrees higher temps than the 2011 Volt) see:

http://avt.inel.gov/pdf/fsev/fact2011nissanleaf.pdf" onclick="window.open(this.href);return false;

As discussed on the LEAF thread here:

http://www.mynissanleaf.com/viewtopic.php?f=31&t=13265" onclick="window.open(this.href);return false;

As to the comparing the LEAF/Volt efficiency ratings at different speeds normalized for temperature-have fun arguing!

At a glance, DC efficiency looks to me to be very close at lower speeds, with the Volts pulling ahead as speed increases, which, I believe, is pretty much in line with what most have thought likely.

But the opposing curves in Volt/LEAF overall trip efficiency (and whether that is actually more a reflection of variations in charge allowed by the individual Battery management systems, as well as the LEAF's passive TM) makes the comparison...Complicated...

http://www.mynissanleaf.com/posting.php?mode=quote&f=10&p=364737" onclick="window.open(this.href);return false;

 

[DaveEV]

It's clear to me that the DC packs have increased internal impedance compared to the L2 packs. This agrees with testing of my own leaf, which is quick charged heavily. I have two GID meters, both with which I can measure impedance of my pack independent of the car's systems. Over the last two months I have noticed a sharp uptick in impedance from about 80 mOhm to 100 mOhm at 30 C. I have also noted that driving on the freeway at higher speeds seems to "cost" me more than just what air resistance alone would account for.

Just as a back-of-envelope calculation, 45 mph uses around 10 kW of power to maintain. 100 mOhm and a nominal voltage of 360V would pull around 28A from the pack with internal pack heating of around 80W.

That amount of internal resistance simply can't result in any measurable range loss.

Even at 65-70 mph and pulling around 20 kW (~55A), the pack is only losing 300W total to heat - or about 1.5% of total energy used.

Something else must explain the difference in range between the L2 and DCQC cars on the 45 mph constant speed test.

One theory: Perhaps the internal resistance is significantly higher at very low SOC, thus triggering the 3.0V cutoff sooner? You can see on page 4 of this report that internal resistance does jump at low SOC: http://www1.eere.energy.gov/vehiclesandfuels/avta/pdfs/fsev/battery_leaf_0356.pdf" onclick="window.open(this.href);return false;

I have 28k miles on the car, 53.2 Ah, 64.3 Hx and 81% SOH. 100% Gid count is around 210-215. I've noted my Hx number is is quite a bit lower than others with the same capacity loss.

My car is very similar to yours: 30.5k mi, 52.9 Ah, 63.6 Hx, though my 100% GID is 219-226. QCd less than 10 times, never seen more than 6 temp bars. Most of the 30k miles were in city/urban driving, though. Not many freeway miles.

 

[JeremyW]

Hmm... yeah... I guess it's really not that much loss. Should have borrowed an envelope before posting...

How about this: My 100% charges only go to 88%-91% raw SOC these days. This is L1 or L2 at 16 or 30 amps. Maybe something is stopping the charge early on the DC charged ones that's ultimately caused by higher internal impedance? I'm assuming they are doing two quick charges to get to full since 2011 vintage cars stop at 80% if the session is started below 50% SOC.

 

[bbrowncods]

Is there a procedure to directly measure impedance?

 

[DaveEV]

How about this: My 100% charges only go to 88%-91% raw SOC these days. This is L1 or L2 at 16 or 30 amps. Maybe something is stopping the charge early on the DC charged ones that's ultimately caused by higher internal impedance? I'm assuming they are doing two quick charges to get to full since 2011 vintage cars stop at 80% if the session is started below 50% SOC.

Yeah, if they are charging to 100% on DCQC for the range test that could also explain the difference in constant speed driving tests. I assume they are using lab equipment to charge/discharge the pack for the raw SOC tests.

 

[edatoakrun]

While the range differential between the DC and L2 LEAFs is very interesting, since virtually all LEAFs driven worldwide probably have experienced lower average battery temperatures than even the L2-charged LEAFs in this study have, the DC-charged results may not have much relevance to our capacity and range experiences.

Of more interest (to those LEAF owners who are not using them as taxis in Phoenix) is that the increase in range/capacity, ~5.1% after 50k miles of the L2-charged LEAFs is reported as actually being greater than the decrease of range/capacity, ~4.5% in the DC-charged LEAFs.

DC Fast Charger Use, Fees, Battery
Impacts and Temperature Impacts
on Charge Rates - EV Roadmap

http://avt.inl.gov/pdf/prog_info/DCFCEVRoadmap7PortlandOregonJuly2014.pdf" onclick="window.open(this.href);return false;

DC Fast Charging Impact Study on 2012 Leafs
• Percentage Range and Capacity at 50,000 miles compared to testing
when new

L2 Average DCFC Average
Range 79.0% 69.3%
Capacity 75.2% 72.6%

Increases in vehicle efficiency, due to lower frictional losses from the drivetrain and tires being the most likely suspects, IMO.

 

[edatoakrun]

Pages 8-13 summarize capacity loss data from previous AVTA reports, but I think the charge rate data is new to MNL.

IEA IA-HEV: DC Fast Charger Use,
Fees, Battery Impacts and
Temperature Impacts on Charge
Rates
Jim Francfort
Issues Related to the Fast Charging of Batteries in
Plug-in Electric Vehicles
IEA Implementing Agreement on Hybrid and
Electric Vehicles (IA-HEV) – Nice, France
September 22 & 23, 2014

http://avt.inl.gov/pdf/prog_info/IEA_DCFCImpactStudySept2014.pdf" onclick="window.open(this.href);return false;

 

[DaveEV]

Anyone else find their claims that "2.6% capacity difference @ 50k miles, probably not a significant difference" completely incorrect?

I mean, if you look at the charge on page 9 - it seems pretty clear that the QC LEAFs have the same capacity at 40k miles as the L2 LEAFs at 50k miles. The QC LEAFs are losing capacity 20% faster than the L2 LEAFs.

Now, in real life what does this mean? Unless you do a significant amount of your charging on QC (significant meaning say 20% or more), it's not likely to have a significant effect on the long term rate of capacity loss on your LEAF.

 

[edatoakrun]

Anyone else find their claims that "2.6% capacity difference @ 50k miles, probably not a significant difference" completely incorrect?...

Statistical significance is the low probability of obtaining at least as extreme results given that the null hypothesis is true.[1][2][3][4][5][6][7] It is an integral part of statistical hypothesis testing where it helps investigators to decide if a null hypothesis can be rejected.[8][9] In any experiment or observation that involves drawing a sample from a population, there is always the possibility that an observed effect would have occurred due to sampling error alone.[10][11] But if the probability of obtaining at least as extreme result (large difference between two or more sample means), given the null hypothesis is true, is less than a pre-determined threshold (e.g. 5% chance), then an investigator can conclude that the observed effect actually reflects the characteristics of the population rather than just sampling error.[8]...

http://en.wikipedia.org/wiki/Statistical_significance" onclick="window.open(this.href);return false;

Yes, at first glance it seems significant.

But with only four vehicles under observation, the possibility of sampling error is considerable.

In other words, the study reporter apparently believes there is a significant possibility that both of the DC charged battery packs had other characteristics that lead to more rapid capacity loss.

Since the four LEAFs were otherwise used ~identically, I suppose they think they could have randomly selected the two less durable batteries for the DC treatment, and they might have lost capacity just about as fast with L2 charging.

I expect this will be explained further when the actual study is released.

 

[DesertSprings]

Were the 70k results ever posted? They should be complete by now shouldn't they?

 

[edatoakrun]

More data from the Phoenix torture test, showing the effects on four LEAFs, of using L2 compared to DC charging, by cycling the full available battery capacity twice-daily from "100%" to shutdown (?) for more than 63 k miles, over ~ 26 months:

http://avt.inl.gov/pdf/energystorage/DCFC_Study_FactSheet_EOT.pdf" onclick="window.open(this.href);return false;

Vehicle mileage accumulation was stopped on 12/13/2014 with each vehicle having accumulated about 63,000 test miles. Exact mileage readings at end-of-test were: VIN 1011:63,479. VIN 4582: 63,207. VIN 2183:63,211. VIN 2078:63,406 miles.

Range results have been posted for 60 and 70 mph constant speed tests at 50 k miles.

At 45 mph CS, the results are also shown after ~63 k miles.

~62 % average of the new range for the L2 LEAFs, and ~57% average for the DC LEAFs.

Capacity loss is not reported after either 60 k or 63 k, apparently since some of the cells are now fully cooked:

...The battery packs could not be charged and discharged to the same voltage limits of previous tests due to cell imbalance, thus the absolute capacity cannot be compared to baseline...

The best news (IMO) is that it looks like the full paper will be available shortly:

...A paper titled Effects of Electric Vehicle Fast Charging on Battery Life and Vehicle Performance will be presented at SAE World Congress on April 22, 2015. This paper will be available on the AVTA website the following week.

edit: I found a more precise statement of observed average capacity loss at 50 k miles in note p. 7, here:

@50k 24.6% Capacity Loss vs 27.6% Capacity Loss ~3miles range difference between DCFC, AC L2

http://energy.gov/sites/prod/files/2014/07/f18/vss131_francfort_2014_o.pdf" onclick="window.open(this.href);return false;

 

[DaveEV]

More data from the Phoenix torture test,

The most interesting discrepancy is that the C/3 test is showing about a 25% loss in capacity, but the actual range tests show a 40% loss of range even at a relatively slow speed of 45 mph.

Is the LBC is going into turtle at a higher voltage than they are discharging the packs in their testing? Or is the increase in internal resistance causing the LBC to cut out earlier? Power draw at 45 mph isn't much, though - just under 10 kW.

Also, I find the difference in range from their 2011 LEAF 45 mph test interesting: only 86 miles - but then they only got 17.7 kWh out of the pack during that test. Also interesting that they replaced the pack on that car under warranty (though not surprising if the car was in Arizona).

 

[RegGuheert]

More data from the Phoenix torture test,

The most interesting discrepancy is that the C/3 test is showing about a 25% loss in capacity, but the actual range tests show a 40% loss of range even at a relatively slow speed of 45 mph.

I wouldn't call this a "discrepancy" since I have been predicting (repeatedly) we would lose range faster than we lose capacity for nearly three years:

I think we should all plan on losing drivable range faster than we lose capacity in a BEV.

I even stressed this point with Andy Palmer and other executives at Nissan in Yokohama since they insisted on equating "capacity degradation" and "range degradation".

Is the LBC is going into turtle at a higher voltage than they are discharging the packs in their testing?

Perhaps. Both you and I have noted that LBW has occurred at higher SOCs than it should based solely on capacity changes. it is reasonable to expect similar effects with turtle.

A similar change occurs at the top end of the range. I have noted in the recent past that the cells quickly (within 0.5 miles) drop from near 4.1V to around 4.0V. INL notes a similar effect in stating that they could not get above 390V with their packs at 63,000 miles. This is almost certainly primarily due to resistance effects. The highest-resistance cell pairs will tend to reach the cutoff voltage first, but their SOC will be lower than it would have been when new due to the additional series resistance. Then, during discharge, that same resistance causes the cell voltage to fall quickly below where it would have been for a cell at the same SOC. What I observe as a result is that I used to be able to drive over 10 miles on the first charge bar, now it is difficult to get over seven.

I do think this effect can be reduced (but not eliminated) by charging more slowly at the end of the charge by either using L1 or by running the climate control.

Or is the increase in internal resistance causing the LBC to cut out earlier?

Almost certainly it does. You have noted that total pack resistance has not increased too much, but the only thing that really matters is the resistance of individual cell pairs. All the literature I have read indicates a very wide spread between different cells when it comes to resistance changes over life. This agrees with my observation that it is more difficult to keep the pack balanced as the it ages, particularly in cold weather.

Power draw at 45 mph isn't much, though - just under 10 kW.

But it is more than twice the rate of a C/3 discharge test. A 58-mile drive at 45 MPH discharges the pack at around C/1.33. So we should expect different capacities between the two tests.

BTW, thanks, Ed, for keeping us apprised of this teasing and the data reports. It sounds like the final paper will be on their website in a week or so.

 

[edatoakrun]

More data from the Phoenix torture test,

The most interesting discrepancy is that the C/3 test is showing about a 25% loss in capacity, but the actual range tests show a 40% loss of range even at a relatively slow speed of 45 mph....

It looks to me like either you and RegGuheert may have misinterpreted the data, or else I have.

It seems fairly clear that AVTA is reporting a capacity loss from baseline (~new) of 27.6% (DC) and 24.6% AC L2, after 50k miles.

And as discussed previously, the range loss at 50,000 miles at 45 mph constant speed exceeded capacity loss in the DC LEAFs (~31%) and the range loss for the AC LEAFs, which only lost ~21% of baseline range, was considerably less than the capacity loss.

Exactly how do you conclude "the C/3 test is showing about a 25% loss in capacity"? at ~63 k miles?

How could the packs have gained average capacity, when driven over ~13,000 additional miles of extreme cycling, including another brutal Phoenix Summer?

BTW, I ran my first range/capacity test of 2015 under warm conditions last Sunday, ~103.4 miles, 204 gids ("100%" charge) to 23 gids, and the loss of range at relatively low speeds occurring at a lower rate than loss of actual available capacity, (and at a much lower rate than the rate of gid loss) is now very obvious in my LEAF, after ~35k miles.

 

[RegGuheert]

It looks to me like either you and RegGuheert may have misinterpreted the data, or else I have.

You're right: I have misinterpreted the data. Thanks for the correction!

In fact, the report indicates the DCFC vehicles lose more range than capacity while the L2 vehicles lose LESS range than capacity:
- At 50,000 miles, the DCFC vehicles lost 27.3% of their capacity but lost 31.4% of their range.
- At 50,000 miles, the L2 vehicles lost 24.7% of their capacity but only 21.1% of their range.

I guess I wonder how this is even possible. Tires may have some impact. Perhaps the batteries weren't well balanced for the initial capacity test?

But the effect reported by INL at 63,000 miles is real. It is harder to get the batteries full as they age and no longer match each other as well. Even at lower levels of degradation, I see this clearly in the very cold weather when resistance is more significant.

OTOH, I can still achieve significant range when the weather is warm if I am careful to fully balance and charge the cells, but it is definitely harder to achieve this than in the past.

 

[edatoakrun]

The paper is now available at the link below.

Effects of Electric Vehicle Fast Charging on Battery Life and Vehicle Performance

http://avt.inl.gov/pdf/energystorage/FastChargeEffects.pdf" onclick="window.open(this.href);return false;

Summary
The four BEVs driven in Phoenix, Arizona were faced with a hot climate, and two were fast charged twice as often as recommended by their manufacturer. Despite these conditions, the vehicles were operated without failure for 50 thousand miles. A greater loss in battery capacity was observed for the fast charged vehicles, though the difference compared to the level two charged vehicles was small in comparison to the overall capacity loss. The vehicle operation was, as intended, verified to be very similar between test groups, and the largest difference in conditions noted was battery temperature during charging. Hotter ambient temperatures appear to have accelerated capacity loss for all of the vehicles in the study, though the exact relationship remains to be seen. Testing is currently underway for two packs, identical to those tested in this study, under constant temperature conditions and identical cycling in the laboratory. The results of that testing combined with the data presented in this paper will serve to further answer the questions related to the rate of capacity, both in relation to time and temperature and will remove even small variations in conditions between the packs.

After a quick look, answers to a few of the more important questions, and one omission worth noting.

They were not driven till dead, as implied previously but:

...Drivers were instructed to drive with the flow of traffic, and to follow a set route consisting of highway and city portions. The route returned the vehicles to an urban loop for the final part of the drive, keeping them close to the charging facility as the remaining range indicator approached zero. This loop was driven until the dashboard range indicator read 5 miles, upon which the vehicle would return to the nearby charging facility and immediately begin charging...

All of the LEAFs were immediately recharged to "100%" after each drive cycle, meaning most of the means of battery abuse available (and which Nissan advised against in the sales disclosure) were utilized.

The Whr/mile (DC) averages reported ranged from 225 to 231 for the four LEAFs, allowing calculation of total kWh throughput.

Gid/battery capacity bars losses don't seem to be mentioned.

I suspect the authors, who did see those LBC reports, would probably get a good laugh from any suggestion to rely on either as a data source.

edit:

On page 12 there is a statement indicating Nissan's claims of relatively rapid initial capacity loss, and a slower rate of loss later in the pack life may be correct, as seen in the preliminary results reported from other packs being tested at constant temperatures:

...Testing and cycling of identical packs in a fixed temperature chamber, in-progress at the writing of this paper, show an initial rate of capacity loss that begins steeply, but slows for each successive test interval approaching a constant rate of capacity loss. The full results of that testing will be presented in a future publication and will serve to answer some questions posed by this testing and analysis...

 

[GRA]

Thanks for providing the link to this. What's unfortunately been missing from this test is any comparison to a group of vehicles operating in a temperate climate and undergoing the same cycling, so that we would have some idea of the point at which battery cooling starts to make a big difference. All we know at the moment is that in Phoenix temps, the additional loss of capacity due to a couple of widely spaced QCs per day is less important than the ambient temp. We also still don't know what the effect of frequent QCing (as advocated by those who want to see QCs spaced every 50 miles along freeways) would be, although we do know that such treatment in an uncooled pack like the LEAF's will boost the battery temp well up into the danger zone.

 

[RegGuheert]

edit:

On page 12 there is a statement indicating Nissan's claims of relatively rapid initial capacity loss, and a slower rate of loss later in the pack life may be correct, as seen in the preliminary results reported from other packs being tested at constant temperatures:

...Testing and cycling of identical packs in a fixed temperature chamber, in-progress at the writing of this paper, show an initial rate of capacity loss that begins steeply, but slows for each successive test interval approaching a constant rate of capacity loss. The full results of that testing will be presented in a future publication and will serve to answer some questions posed by this testing and analysis...

Of course they get the same result as Nissan: they're doing the same test which is (likely) dominated by cycling losses. Cycling losses slow as the battery ages.

But with very few exceptions, losses customers experience are dominated by calendar losses, which tend to be linear or even extralinear with time.

It sure would be nice if someone would do a calendar aging test in parallel with these cycling tests, since that is the dominant loss factor for most LEAF owners.