Capacity Loss on 2011-2012 LEAFs

[Yanquetino]

If an owner wants to estimate his/her likely remaining capacity at a given point in time, the Battery Aging Model is the most accurate predictor we currently have.

I hope Nissan endorses it, then! Indeed, if they do deem it accurate and reliable, I hope that they compensate you for the effort and start to publish a variation of it on their website.

 

[Stoaty]

If an owner wants to estimate his/her likely remaining capacity at a given point in time, the Battery Aging Model is the most accurate predictor we currently have.

I hope Nissan endorses it, then! Indeed, if they do deem it accurate and reliable, I hope that they compensate you for the effort and start to publish a variation of it on their website.

We are all hoping that Nissan will create a tool similar to the Battery Aging Model (they certainly have tons more data than we do) and make it available to help people decide whether a Leaf is right for them. I would like to see them make this information available in simplified form and communicated to prospective purchasers at the time of sale/lease. Then those that live in Phoenix would know that perhaps a 2 years lease would be the best for them (or maybe they should skip the Leaf altogether), while in Seattle a Leaf purchase is a slam dunk. The Leaf is a great car, but if Nissan doesn't disclose realistic expectations about both range and battery capacity prior to sale they are asking for trouble later on.

PS The Battery Aging Model predicts 90.5% remaining capacity for me after 17 months and 12,400 miles. From my Gid meter reading (90.7% on full charge) it appears to be spot on. EOL for me is somewhere between 9.5 years (Van Nuys, near work) and 11.4 years (Santa Monica, near home), so the Leaf will work out great for me if my model is correct.

 

[TonyWilliams]

If an owner wants to estimate his/her likely remaining capacity at a given point in time, the Battery Aging Model is the most accurate predictor we currently have.

I hope Nissan endorses it, then! Indeed, if they do deem it accurate and reliable, I hope that they compensate you for the effort and start to publish a variation of it on their website.

I personally find it hard to believe that we will educate Nissan on anything. If they had published this type of data on day 1 of sales, it would be a hard sell. Everybody would have demanded a TMS.

So, why would they acquiesce now? They've sold 40,000 cars without this data. I doubt this type of information will sell more cars, and quite possibly quell future sales. They've made it bruised a bit this summer, and next summer will be a slaughter, but I think the average Joe SixPack won't have a clue or care. If it's cheap enough, they'll buy it.

 

[jlsoaz]

Agreed with Tom. We got a link to the thread. The actual survey link appears to be here:

http://www.pluginamerica.org/surveys/batteries/leaf/survey.php" onclick="window.open(this.href);return false;

I filled it out.

At the end, the results are posted, which I bookmarked as here:

http://www.pluginamerica.org/surveys/batteries/leaf/results.php" onclick="window.open(this.href);return false;

As to getting good data on miles driven, I don't know that some generalized (not-Leaf-specific) number is going to give us what we need, whether for AZ or another place. It seems useful to know I guess in some ways, such as comparing # of miles driven in each state for cars in general as versus sold and leased leafs (separating them since there is a clear financial incentive to keep leased leaf # of miles below the cutoff).

The points made about questioning the trustworthiness of Carwings data seem worth considering. (I, by the way, so far fall into "hit whatever the hell I feel like at that moment Decline/ or Ok/". I regard the problem of the ok nag as both a business and safety-related nuisance.)

There is the issue of assessing whether the 7,500 is accurate. Even if it turns out to be accurate, it sounded well worth questioning in my view as to how it was used.

It might be worthwhile listing the link to the survey. I, for one, did not even know that they currently had one underway...

which makes the pluginamerica.org survey important to fill out. we need data and AZ seems to be one of the areas we are not getting a lot of participation in!

 

[DaveEV]

The Lithium-Ion coverage period is 96 months or 100,000 miles, whichever comes first.

In other words, if I have driven 100,000 miles in only 4 years (instead of 8), the warranty expires. Similarly, if after 8 years I have driven only 20,000 miles (instead of 100,000), the warranty expires.

Whichever comes first.

This especially makes sense to me with batteries, because they also deteriorate with age --not just miles.

While this is true in terms of warranty, they are not directly interchangeable.

For example, a LEAF driven 100k miles in 1 year (basically driven around the clock) will probably see much less degradation than a LEAF driven 100k miles in 8 years.

Most of us here on the forum know what will speed up capacity loss - primarily battery temperature, number of cycles and time spent at specific state of charge (more time at higher SOC = faster capacity loss).

The problem we all have is that Nissan has only given us 2 time/mile numbers to estimate capacity loss - 5 years / ~60k mi at 80% and 10 years / 120k mi at 70% capacity remaining while in reality based on stoaty's excellent battery model, Arizona customers will probably see that type of capacity loss twice as fast and Seattle customers will probably see that type of capacity loss half as fast. Until Nissan specifically comes out and specifically says this, the typical customer being optimistic is going to look at Nissan's average estimate and assume that if they treat the battery well they're going to see capacity loss at very similar rates regardless of where they live.

 

[Yanquetino]

We are all hoping that Nissan will create a tool similar to the Battery Aging Model (they certainly have tons more data than we do) and make it available to help people decide whether a Leaf is right for them. I would like to see them make this information available in simplified form and communicated to prospective purchasers at the time of sale/lease. Then those that live in Phoenix would know that perhaps a 2 years lease would be the best for them (or maybe they should skip the Leaf altogether), while in Seattle a Leaf purchase is a slam dunk. The Leaf is a great car, but if Nissan doesn't disclose realistic expectations about both range and battery capacity prior to sale they are asking for trouble later on.

PS The Battery Aging Model predicts 90.5% remaining capacity for me after 17 months and 12,400 miles. From my Gid meter reading (90.7% on full charge) it appears to be spot on. EOL for me is somewhere between 9.5 years (Van Nuys, near work) and 11.4 years (Santa Monica, near home), so the Leaf will work out great for me if my model is correct.

Hi, Stoaty:

I applied the data from your table for Phoenix to plot its corresponding polynomial curve for battery loss over time. Compared to my own graph that uses Nissan's generic 5- and 10-year benchmarks, It differs significantly, as you can see:

If your model is correct, the 80% and 70% capacity losses certainly do occur much earlier, because your proposed 1.81 aging factor for Phoenix drops the curve much lower and at a faster rate. Indeed, the model projects a 70% capacity several months before Nissan anticipates even 80%.

I am still puzzled, however, when I see how your model's curve applies to the ranges achieved in the owners' test. To compare projections, I have overlaid your model's polynomial curve on top of my own graph of those ranges. We can see that your model's purple curve does indeed plot lower and drops faster than Nissan's green curve:

As for the ranges achieved in the test, the 2 red outliers are still notably below your curve; 1 Leaf is barely below it; 2 are on top of it; and... 7 Leafs are above it. It therefore seems to me that, despite what their faulty gauges were predicting, most of the Leafs tested actually achieved better ranges than one might expect in Phoenix using your Battery Aging Model...?

As you can see, these are the same 7 Leafs, curiously enough, that fell within their corresponding error bars of Nissan's projected ranges in my analysis.

I dunno. I think we really do need Nissan to come up with some definitive data to put all the speculations to rest.

 

[Stoaty]

I am still puzzled, however, when I see how your model's curve applies to the ranges achieved in the owners' test. To compare projections, I have overlaid your model's polynomial curve on top of my own graph of those ranges. We can see that your model's purple curve does indeed plot lower and drops faster than Nissan's green curve:

My battery aging model cannot be applied to most of the Leafs in the Arizona range test, because I do not know the number of months each car was owned at the time of the test (in other words, I need to know the delivery date for each car). The model uses the city (to get the climate aging factor), years of ownership (probably needs to be accurate to within one month to provide meaningful results), and total miles driven to predict remaining capacity. If Tony can supply this information, I will do the calculations. The length of ownership is very important, because battery capacity drops rapidly due to calendar losses in the first 12-18 months.

Edit: actually, to fully apply the model I also need to know the average number of days (or fraction thereof) that the car was parked in the sun. That might be a pretty significant factor in Phoenix.

 

[Yanquetino]

My battery aging model cannot be applied to most of the Leafs in the Arizona range test, because I do not know the number of months each car was owned at the time of the test (in other words, I need to know the delivery date for each car). The model uses the city (to get the climate aging factor), years of ownership (probably needs to be accurate to within one month to provide meaningful results), and total miles driven to predict remaining capacity. If Tony can supply this information, I will do the calculations. The length of ownership is very important, because battery capacity drops rapidly due to calendar losses in the first 12-18 months.

Edit: actually, to fully apply the model I also need to know the average number of days (or fraction thereof) that the car was parked in the sun. That might be a pretty significant factor in Phoenix.

Yeah, that makes a lot of sense. I'm sure that is a significant factor too. And then there are soooo many other factors, depending upon "care and usage." There simply must be other reasons for those few outliers besides age, miles, and climate --or else there would be more outliers.

 

[TonyWilliams]

There simply must be other reasons for those few outliers besides age, miles, and climate --or else there would be more outliers.

Yes, there are. Parked in a garage while charging in Phoenix with a west facing door (and hit with the full power of the sun).

That garage had to be 130F-150F. That was the 29,000 mile car that drove 59 miles in our test.

 

[TonyWilliams]

If Tony can supply this information, I will do the calculations. The length of ownership is very important, because battery capacity drops rapidly due to calendar losses in the first 12-18 months.

Edit: actually, to fully apply the model I also need to know the average number of days (or fraction thereof) that the car was parked in the sun. That might be a pretty significant factor in Phoenix.

Car -------Miles ------Manufacture
Red429 - 11,500 ----- 3/2011
Blue494 - 29,000 ----- 4/2011
Blue534 - 16,000 ----- 4/2011
White530-12,000 ----- 4/2011
White272-17,500 ----- 3/2011
Red500 - 22,500 ----- 2/2011
White626-17,300 ----- 4/2011
Blue842 -- 2,500 ----- 4/2012
Silver679-14,750 ----- 5/2011
Blue917 - 13,900 ----- 5/2011
Black782--7,000 ----- 4/2012 (San Diego)
Blue744- 22,400 ----- 4/2011
RedXXX -- 100 ------ 8/2012

 

[GerryAZ]

Stoaty,
Here is the age data for my car (Silver 679 in the range test): I purchased it on June 12, 2011, it was manufactured in May 2011, and I have been averaging slightly over 1,000 miles per month. It was delivered to the dealer late on Friday, they prepared it on Saturday, and I drove it home on Sunday morning so it did not sit on the dealer's lot long. The odometer reads 17349 right now and I have only used quick charging twice. My car is parked in the shade at work, at the airport, and in my driveway so it is only parked in the sun while I am shopping or going to restaurants, etc. Although I always charge to 100% using my L2 AeroVironment EVSE because I need the range, the SOC is usually at about 6 bars when I park at the office or airport. If you need any more information to plug into your model, please ask.

Yanquetino,
Tell me what number I should use for the actual energy from the wall for Level 2 charging of a new LEAF battery from shutdown (dead) to 100% and I will tell you what my actual battery capacity loss is when it finishes charging in the morning.

Gerry

 

[Stoaty]

Stoaty,
Here is the age data for my car (Silver 679 in the range test): I purchased it on June 12, 2011, it was manufactured in May 2011, and I have been averaging slightly over 1,000 miles per month. It was delivered to the dealer late on Friday, they prepared it on Saturday, and I drove it home on Sunday morning so it did not sit on the dealer's lot long. The odometer reads 17349 right now and I have only used quick charging twice. My car is parked in the shade at work, at the airport, and in my driveway so it is only parked in the sun while I am shopping or going to restaurants, etc.

I forgot to mention one other thing I need:

Average miles/kwh from the dash over life of your Leaf.

PS I used an average of 4.0 and got 86.5% remaining capacity predicted compared to 84.5% actual in the range test, but having the actual number may make the prediction more (or less) accurate.

 

[turbo2ltr]

New record, which is above the point I lost the first bar, yet I am still 2 bars down. I don't think capacity bars will ever comeback without intervention from nissan.

Also of note,we've had unseasonably cool temps. Lows in the 40s.

GIDs at 100% charge

5/2012 (one bar lost): 228
6/30/2012 (two bars lost): 213
7/22/2012: 215
9/21/2012: 213
10/23/2012: 217
11/5/2012: 220
11/12/2012: 230

But I noticed something weird last night. I was down 6 SOC bars and still didn't have all the regen bubbles. The last bubble would come and go..it was rather odd. I've seen this at very high soc, but not at 50%.

Cell balance is consistently around 40-45mv difference.

 

[Stoaty]

Car -------Miles ------Manufacture
Red429 - 11,500 ----- 3/2011
Blue494 - 29,000 ----- 4/2011
Blue534 - 16,000 ----- 4/2011
White530-12,000 ----- 4/2011
White272-17,500 ----- 3/2011
Red500 - 22,500 ----- 2/2011
White626-17,300 ----- 4/2011
Blue842 -- 2,500 ----- 4/2012
Silver679-14,750 ----- 5/2011
Blue917 - 13,900 ----- 5/2011
Black782--7,000 ----- 4/2012 (San Diego)
Blue744- 22,400 ----- 4/2011
RedXXX -- 100 ------ 8/2012

Helpful, but the model uses date of purchase rather than date of manufacture. I can put in a guesstimate of 2 months from manufacture to delivery, but it would be better to have actual date of purchase, since that was used to tune the model. If you can PM me the owner names (forum handle) for each car, I can probably get most of the information from the Wiki. I have been meaning to work on the validation of the model, but haven't gotten around to it until now.

 

[surfingslovak]

The last bubble would come and go..it was rather odd. I've seen this at very high soc, but not at 50%.

Thank you for the report! The regen bubble behavior could be temperature-related. Do you garage your car or park it outside? How cold was the battery at the time you made this observation?

The length of ownership is very important, because battery capacity drops rapidly due to calendar losses in the first 12-18 months.

Stoaty, this is fantastic! On a related note, I worked with ElectricVehicle and FalconFour to get a Gid count reading for his Leaf. If you recall, this car is another outlier, as it sat on dealer lots in Bakersfield and Fresno for a year. FalconFour only owned it for couple of months and put 2,000 miles on it. The car finished a full charge with 258 Gids. We let it sit for couple of hours and then restarted the charging session. It charged for another 20-25 minutes, and the Gid count went to 263. Pack voltage was 394.5. I'd think that FalconFour will post photos later this week. He drove 360 miles round-trip from Fresno to help us with this data point.

 

[DaveinOlyWA]

New record, which is above the point I lost the first bar, yet I am still 2 bars down. I don't think capacity bars will ever comeback without intervention from nissan.

Also of note,we've had unseasonably cool temps. Lows in the 40s.

GIDs at 100% charge

5/2012 (one bar lost): 228
6/30/2012 (two bars lost): 213
7/22/2012: 215
9/21/2012: 213
10/23/2012: 217
11/5/2012: 220
11/12/2012: 230

But I noticed something weird last night. I was down 6 SOC bars and still didn't have all the regen bubbles. The last bubble would come and go..it was rather odd. I've seen this at very high soc, but not at 50%.

Cell balance is consistently around 40-45mv difference.

the regen bubble thing i have seen as well. i wrote it off as some sort of control sensor taking its time to settle down. I notice this MUCH more often in the first 5 minutes after start up and it doesnt always happen right away. sometimes circles will start popping off after 2 minutes or something.

see it a lot after a QC session too. a few times, i would charge to only 75-80% SOC and have NO regen circles for first 3-5 minutes of driving

 

[Yanquetino]

Yanquetino,
Tell me what number I should use for the actual energy from the wall for Level 2 charging of a new LEAF battery from shutdown (dead) to 100% and I will tell you what my actual battery capacity loss is when it finishes charging in the morning.

Gerry

Hi, Gerry:

Given his thorough knowledge of the Leaf's inner workings, I bet Phil (Ingineer) could provide a reliable answer to this question. I can only venture wild guesses.

Let's see. The Leaf is advertised as having a 24 kW pack, but supposedly only 21 kW are useable. Then there is the efficiency loss from the EVSE to the car, which I think most would agree is around 85%. With those caveats, my wild guess is that it would take about 24.7 kWh from the wall to fill a spanking new Leaf from "empty" to "full" capacity. Nissan's parameters would suggest that it should take about 23.2 kWh from the wall to "fully" charge a Leaf after 15 months/15,000 miles. From there, my next wild guess is that it might take about 22.2 kWh from the wall to "fill" your Leaf, since it tested toward the lower end of its corresponding scale of ranges. Do you have a meter on your EVSE to measure this? That would be useful!

 

[surfingslovak]

Yanquetino,
Tell me what number I should use for the actual energy from the wall for Level 2 charging of a new LEAF battery from shutdown (dead) to 100% and I will tell you what my actual battery capacity loss is when it finishes charging in the morning.

Given his thorough knowledge of the Leaf's inner workings, I bet Phil (Ingineer) could provide a reliable answer to this question. I can only venture wild guesses.

Let's see. The Leaf is advertised as having a 24 kW pack, but supposedly only 21 kW are useable. Then there is the efficiency loss from the EVSE to the car, which I think most would agree is around 85%. With those caveats, my wild guess is that it would take about 24.7 kWh from the wall to fill a spanking new Leaf from "empty" to "full" capacity. Nissan's parameters would suggest that it should take about 23.2 kWh from the wall to "fully" charge a Leaf after 15 months/15,000 miles. From there, my next wild guess is that it might take about 22.2 kWh from the wall to "fill" your Leaf, since it tested toward the lower end of its corresponding scale of ranges. Do you have a meter on your EVSE to measure this? That would be useful!

Gerry, correct me if I'm wrong, but I thought we discussed this before and you measured 20.13 kWh from the wall on level 2? As to new batteries, the EPA test performed by Nissan on a factory-new vehicle implies about 24.5 kWh, depending on the cycle used, which was discussed on the forum before. That's at least what's baked into the official MPGe figures on the Monroney sticker. The NREL tear-down report dress found a while ago, has 25.41 kWh from the wall. It also recorded charging efficiencies and measured the usable energy output from a new battery at 21.38 kWh.

None of this came as a surprise, and we collectively took it as confirmation of previous work done here and elsewhere. Keep in mind that Yanquentino's analysis and charts do not reflect or consider any climatic influences at all. The numbers and recommendations his website provides apply to all vehicles all the same, regardless of their geographic location. He goes strictly by the book, Nissan's book that is, and does not consider any empirical data outside of the Phoenix range test in his analysis. I hope this helps.

 

[Yanquetino]

Keep in mind that Yanquentino's analysis and charts do not reflect or consider any climatic influences at all. The numbers and recommendations his website provides apply to all vehicles all the same, regardless of location. He goes strictly by the book, Nissan's book that is, and does not consider any empirical data outside of the Phoenix range test in his analysis. I hope this helps.

Mmm. Not entirely true, George. Yes, my figures are based upon "Nissan's book" to determine what the automaker's parameters would predict. I do not think, however, that said "book" fails to "consider any climatic influences at all." I think it reasonable to assume that this might be one of the reasons why Nissan postulates a range of ranges (76-to-84 miles), rather than just the average (80 miles), illustrated with the error bars in my chart. I might be wrong, but I think the automaker anticipates that, even with "normal operation and recommended care," factors such as heat and cold can --and will-- affect range to some degree. After all, they explicitly state such differences in their 6 scenarios, so I think my assumption has merit.

Whether or not that assumption is correct is probably something that the advisory board can ask Andy Palmer to clarify.

 

[Stoaty]

Here are preliminary results. I estimated 2 months from manufacture to delivery. Does not include efficiency (miles per kwh from dash) or solar loading for most of the cars listed.