Suggest Cities to See Relative Battery Aging Factor

[Weatherman]

The last five on my list:

Denver - 0.70
Chicago - 0.75
LAX - 0.89
Nashville - 1.02
Van Nuys - 1.08

Hopefully, that should be enough independent confirmation of the weatherspark results.

 

[Stoaty]

The last five on my list:

Denver - 0.70
Chicago - 0.75
LAX - 0.89
Nashville - 1.02
Van Nuys - 1.08

Hopefully, that should be enough independent confirmation of the weatherspark results.

Added to Wiki, thanks. The results are very close with the exception of a band in the high end around 1.5 where a few of the cities are about 5% lower with your numbers--not enough to make much difference when calculating battery degradation.

 

[Stoaty]

Stoaty, it's great to see all this progress, and I would be happy to elaborate. I based the model on couple of papers we discussed earlier, and a handful of control cars. The model could be definitely be refined, but we'll need more data.

One interesting reference is FalconFour (see above). His Leaf essentially sat on a parking lot for a year, and has negligible cycling losses.

Aside from that, I used two Leafs from Casa Grande, TickTock's and azdre's, since they both had battery health figures available. Then Steve Marsh, Tony, my friend's car with 9% dealer-measured loss, and my own vehicle.

Perhaps it would make sense to tabulate these data points and make sure that the model conforms as we add more values.

Thanks, I probably forgot the discussion, or didn't fully understand it. I assume you are going to update the model with our aging factors relative to the Civic Center. Did you include the data from Nissan Normal (TM) for a Leaf driving 12500 miles in L.A.? How about the 76% capacity retention for Phoenix Leafs driving 7500 miles per year? The model should probably be tweaked so that it gives the right answer for those two cases, although for the L.A. case that isn't quite the same as actual data from cars driven.

Thanks for your pioneering efforts to develop the model and get the data. I only jumped on board when I realized my Photoshop skills could come in handy.

 

[DaveinOlyWA]

One interesting reference is FalconFour (see above). His Leaf essentially sat on a parking lot for a year, and has negligible cycling losses.

that is probably not correct. he stated he ran his LEAF down to the point where he had to push it into his parking lot to charge and put in 25Kwh per Kill a watt on L1.

with an estimated 76% efficiency rate, only 19 Kwh would have made it to the battery suggesting he has lost 10%

 

[Stoaty]

One interesting reference is FalconFour (see above). His Leaf essentially sat on a parking lot for a year, and has negligible cycling losses.

that is probably not correct. he stated he ran his LEAF down to the point where he had to push it into his parking lot to charge and put in 25Kwh per Kill a watt on L1.

with an estimated 76% efficiency rate, only 19 Kwh would have made it to the battery suggesting he has lost 10%

I think you just confirmed surfingslovaks statement. He said negligible cycling losses and estimated 8-10% loss of capacity from calendar loss.

 

[DaveinOlyWA]

One interesting reference is FalconFour (see above). His Leaf essentially sat on a parking lot for a year, and has negligible cycling losses.

that is probably not correct. he stated he ran his LEAF down to the point where he had to push it into his parking lot to charge and put in 25Kwh per Kill a watt on L1.

with an estimated 76% efficiency rate, only 19 Kwh would have made it to the battery suggesting he has lost 10%

I think you just confirmed surfingslovaks statement. He said negligible cycling losses and estimated 8-10% loss of capacity from calendar loss.

oh ok gotcha ya. i took his quote from your post. i have to say that calendar loss in his case can only partially be attributed. Fresno is a pretty warm place right? heat is also probably a contributing factor as well

 

[Weatherman]

Added to Wiki, thanks. The results are very close with the exception of a band in the high end around 1.5 where a few of the cities are about 5% lower with your numbers--not enough to make much difference when calculating battery degradation.

It may be possible that the hourly temperatures at some of the tropical and humid subtropical locations are moderated by afternoon clouds and rain during the warmer part of the year. This may account for the lower numbers. But, as you said, the difference between a 1.59 and a 1.69 isn't going to make much difference.

 

[FalconFour]

oh ok gotcha ya. i took his quote from your post. i have to say that calendar loss in his case can only partially be attributed. Fresno is a pretty warm place right? heat is also probably a contributing factor as well

It's pretty warm - gets up to 100-110F for a few weeks out of the year - maybe 4 or 5 weeks of those temperatures year-to-date. Rest of the time it's a pretty wide range of temperatures 80-90, a few weeks of 60-80, then a steep drop to 40-60, but only a week of near-freezing (30-40) temps - never snow, occasional rain.

If I could just try to remember what months in the year correspond to what temperatures so I'd quit getting so surprised by the weather even though I've lived here all my life... maybe I could consider myself "grown up" :lol:

 

[RegGuheert]

I just took a look at the Wiki page on this and I must say this is AWESOME!! Kudos to Stoaty, surfingslovak and Weatherman for putting this together!!! (Also kudos to everyone else who has contributed!) It's really good to have some sort of educated guess about where we will come out in terms of capacity losses based on our climate. You guys even included a city close to me!

One question: Does anyone have a guess about what city in that list matches Nissan's idea of the worldwide median (or mean) for capacity loss? Perhaps it is someplace cooler than LA...

One request for surfingslovak: Can you please adjust your algorithm so that I will think our LEAF's battery will last as long as the ones in Seattle? :lol:

 

[Weatherman]

One question: Does anyone have a guess about what city in that list matches Nissan's idea of the worldwide median (or mean) for capacity loss? Perhaps it is someplace cooler than LA...

Well, the current numbers assume downtown L.A. is the "standard". That's why it's 1.0. Any location with a number greater than 1.0 would expect to have higher degradation rates than Nissan Normal (TM), and any location with a number less than 1.0 would expect to have lower degradation rates than Nissan Normal (TM). All other things being equal, of course.

 

[RegGuheert]

Well, the current numbers assume downtown L.A. is the "standard". That's why it's 1.0.

Thanks. I get that the table on the Wiki is normalized around LA. But I am under the impression that LEAF batteries in LA are degrading FASTER than Nissan's worldwide median car.

 

[Weatherman]

The challenge of using downtown L.A. as the standard is that temperatures can vary by such a wide margin just within the city limits. On a summer afternoon, it can be 65 at the beach, while it's close to 100 inland. That's why I thought Nashville would be a better choice. Nashville comes in close to downtown L.A. on the scale, and there's much less variability in the local area. Tennessee residents can correct me if I'm wrong, but I don't recall any of them seeing a capacity bar loss, yet.

 

[RegGuheert]

It's just that Andy Palmer made it clear in the video with Chelsea Sexton that the 80% capacity after five years number is a worldwide median. Since more LEAFs have been sold in Japan than in the US, perhaps they use Tokyo as their "median city." I could be way off base, but it just seems that nearly every international city is below LA in the chart in Wiki.

 

[FalconFour]

I should add... my figures may be off. I cited 25kWh as a figure I had seen on my Kill-A-Watt meter, reset a few minutes after starting the charge just for curiosity's sake, from a non-scientific experiment. Since L1 charging takes 22 hours from dead, and I drive the LEAF every day, I didn't plan out an "experiment" (I just have the meter there "watching everything"). It could be that the driving+charging from the Nissan station put more in than I used! I'll have to get another sample if my figure there is going to be of some significance.

 

[Stoaty]

Well, the current numbers assume downtown L.A. is the "standard". That's why it's 1.0.

Thanks. I get that the table on the Wiki is normalized around LA. But I am under the impression that LEAF batteries in LA are degrading FASTER than Nissan's worldwide median car.

Not necessarily. The batteries that are degrading faster are mostly in hotter areas or have higher mileage. Besides, we don't know how (in)accurate the capacity bars are. I am pretty sure that Andy Palmer said in the video that the 80% figure was for cars in a climate "like Los Angeles" for 12500 miles a year. At any rate, with Nashville at 1.02 and Ota, Japan at 0.98 it doesn't make much difference which standard is used--they are the same within our ability to measure the (changing) climate.

 

[Stoaty]

OK, I found some data from the range test in Arizona that supports my hypothesis that cycling losses are currently too low for that environment in surfingslovaks model. Using his spreadsheet Nissan Degradation Estimator (NDE) (TM) we see that in Phoenix he uses 3.02% degradation for every 10,000 miles. However, the range test showed a very high correlation between total miles driven (correlation coefficient 0.84) and percent capacity. Linear regression showed that for every 10,000 miles driven there was a 7.5% loss of capacity (see graph at link below):

http://www.mynissanleaf.com/wiki/index.php?title=Battery_Capacity_Loss#Range_Test_on_Cars_with_Battery_Capacity_Loss" onclick="window.open(this.href);return false;

Presumably calendar losses were roughly the same for all cars (need to check their ages to see how much of a factor that might be), so the extras 7.5% loss of capacity should all be due to cycling losses.

It seems to me that this is the sort of data that should be incorporated into the model (especially since Nissan blamed "high mileage" of the affected Leafs for the rapid capacity loss). Thoughts?

 

[DaveEV]

Thoughts?

Good stuff. I suspect that the correlation data you're coming up with will tighten up if you had accurate production and delivery dates.

 

[Stoaty]

Thoughts?

Good stuff. I suspect that the correlation data you're coming up with will tighten up if you had accurate production and delivery dates.

Agreed. This data would help to accurately estimate calendar losses, and we could assume a more benign aging factor between production and delivery to the U.S., then apply the aging factor appropriate to the local environment if we knew roughly when the Leaf was delivered to the dealer and roughly when it was delivered to the customer. Bottom line: the more data we have about both the Casa Grande leafs and those tested in Arizona by Tony and crew, the more helpful it will be.

 

[surfingslovak]

Thoughts?

Good stuff. I suspect that the correlation data you're coming up with will tighten up if you had accurate production and delivery dates.

I took TickTock's graph quite literally. Probably not fair, but in the absence of better data, it will have to do. Since it shows 8% Normal(TM) degradation after 12 months, I ballparked calendar life losses at 6% and cycling losses around 2% at the reference location. Cross-referenced Tony's chart for that figure as well. I'm fairly confident that cycling losses are nearly linear over the life of the battery. Although they seem to diminish as the pack ages. Likewise, higher temps imply higher cycling losses, but we don't know if the same relative aging factor should be used like for calendar life.

 

[Stoaty]

I took TickTock's graph quite literally. Probably not fair, but in the absence of better data, it will have to do. Since it shows 8% Normal(TM) degradation after 12 months, I ballparked calendar life losses at 6% and cycling losses around 2% at the reference location. Cross-referenced Tony's chart for that figure as well. I'm fairly confident that cycling losses are nearly linear over the life of the battery. Although they seem to diminish as the pack ages. Likewise, higher temps imply higher cycling losses, but we don't know if the same relative aging factor should be used like for calendar life.

Well, one thing would be to put in the new battery aging factors relative to Nissan Normal(TM). The factor of aroud 1.8 is higher that the 1.5 or so you have in the model. Perhaps charging to 100% frequently is a factor too (don't know how many of the Casa Grande or range test Leafs charged to 100% a lot and how many were usually at 80%). However, it seems that the actual data suggests quite a bit higher than 2*1.8 = 3.6% for cycling losses. It appears that this should be corrected by a factor of 12500 miles/7500 miles to get a close number (which would be about 6%, relatively close to 7.5% from the range test). Perhaps you have already applied that correction factor in your model? Why do you not consider the information from the range test valid data to incorporate into the model (if I understand you correctly)?